CN119872358A - Seat adjusting method, device, storage medium, controller, vehicle and product - Google Patents
Seat adjusting method, device, storage medium, controller, vehicle and product Download PDFInfo
- Publication number
- CN119872358A CN119872358A CN202510138454.3A CN202510138454A CN119872358A CN 119872358 A CN119872358 A CN 119872358A CN 202510138454 A CN202510138454 A CN 202510138454A CN 119872358 A CN119872358 A CN 119872358A
- Authority
- CN
- China
- Prior art keywords
- information
- target
- occupant
- key points
- passenger
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60N—SEATS SPECIALLY ADAPTED FOR VEHICLES; VEHICLE PASSENGER ACCOMMODATION NOT OTHERWISE PROVIDED FOR
- B60N2/00—Seats specially adapted for vehicles; Arrangement or mounting of seats in vehicles
- B60N2/002—Seats provided with an occupancy detection means mounted therein or thereon
- B60N2/0021—Seats provided with an occupancy detection means mounted therein or thereon characterised by the type of sensor or measurement
- B60N2/0022—Seats provided with an occupancy detection means mounted therein or thereon characterised by the type of sensor or measurement for sensing anthropometric parameters, e.g. heart rate or body temperature
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60N—SEATS SPECIALLY ADAPTED FOR VEHICLES; VEHICLE PASSENGER ACCOMMODATION NOT OTHERWISE PROVIDED FOR
- B60N2/00—Seats specially adapted for vehicles; Arrangement or mounting of seats in vehicles
- B60N2/02—Seats specially adapted for vehicles; Arrangement or mounting of seats in vehicles the seat or part thereof being movable, e.g. adjustable
- B60N2/0224—Non-manual adjustments, e.g. with electrical operation
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/08—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to drivers or passengers
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2420/00—Indexing codes relating to the type of sensors based on the principle of their operation
- B60W2420/40—Photo, light or radio wave sensitive means, e.g. infrared sensors
- B60W2420/403—Image sensing, e.g. optical camera
Landscapes
- Engineering & Computer Science (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Aviation & Aerospace Engineering (AREA)
- Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Cardiology (AREA)
- General Health & Medical Sciences (AREA)
- Heart & Thoracic Surgery (AREA)
- Seats For Vehicles (AREA)
Abstract
本申请涉及一种座椅调节方法、装置、存储介质、控制器、车辆及产品，通过目标车辆内的摄像头，获取所述目标车辆内的目标乘员的目标图像；基于所述目标图像，确定所述目标乘员上关键点的三维特征信息，并基于所述关键点的三维特征信息，预测所述目标乘员的体姿信息；基于所述目标乘员的体姿信息，对所述目标乘员乘坐的座椅进行调节，使得在降低成本的同时提高座椅调节的准确性。
The present application relates to a seat adjustment method, device, storage medium, controller, vehicle and product. A target image of a target occupant in the target vehicle is obtained through a camera in the target vehicle; based on the target image, three-dimensional feature information of key points on the target occupant is determined, and based on the three-dimensional feature information of the key points, the body posture information of the target occupant is predicted; based on the body posture information of the target occupant, the seat in which the target occupant sits is adjusted, so that the accuracy of seat adjustment is improved while reducing costs.
Description
Technical Field
The application relates to the field of vehicles, in particular to a seat adjusting method, a seat adjusting device, a storage medium, a controller, a vehicle and a product.
Background
With the development of science and technology, vehicles are becoming more popular with users, and functions on vehicles are also becoming more and more popular, for example, vehicles can automatically adjust the position of seats in vehicles so as to provide better riding experience for people in vehicles.
However, the current method of automatically adjusting the position of the seat in the vehicle is either costly or less accurate, and thus, the method of automatically adjusting the position of the seat in the vehicle is to be lifted.
Disclosure of Invention
The embodiment of the application provides a seat adjusting method which can reduce the cost and improve the accuracy at the same time so as to at least partially solve the technical problems.
In order to achieve the above object, according to a first aspect of the present application, there is provided a seat adjusting method including:
acquiring a target image of a target passenger in a target vehicle through a camera in the target vehicle;
Determining three-dimensional characteristic information of key points on the target passenger based on the target image, and predicting body posture information of the target passenger based on the three-dimensional characteristic information of the key points;
and adjusting the seat on which the target passenger sits based on the posture information of the target passenger.
Optionally, the posture information includes body shape information, and predicting the body shape information of the target occupant based on the three-dimensional feature information of the key point includes:
Predicting the volume information of the target passenger based on the three-dimensional feature information of the key points;
based on the volume information, body shape information of the target occupant is predicted.
Optionally, predicting the body type information of the target occupant based on the volume information includes:
predicting weight information of the target occupant based on the volume information;
And predicting body shape information of the target occupant based on the weight information and the three-dimensional feature information of the key points.
Optionally, the predicting the body shape information of the target occupant based on the weight information and the three-dimensional feature information of the key points includes:
Predicting initial body shape information of the target passenger based on the three-dimensional characteristic information of the key points by an artificial engineering method;
And performing adjustment processing on the initial body shape information based on the weight information to obtain body shape information of the target occupant.
Optionally, the key point is a key point of an upper body of the target occupant, the initial body shape information includes initial height information, and the predicting the initial body shape information of the target occupant based on three-dimensional feature information of the key point by an ergonomic method includes:
Predicting the height information of the upper body of the target passenger based on the three-dimensional characteristic information of the key points by an artificial engineering method;
Predicting the height information of the lower body of the target occupant based on the height information of the upper body;
And determining initial height information of the target occupant based on the height information of the upper body and the height information of the lower body.
Optionally, the present embodiment further includes:
determining attribute information of the target occupant based on the target image;
The adjusting the initial body shape information based on the weight information to obtain body shape information of the target occupant includes:
And performing adjustment processing on the initial body shape information based on the weight information and the attribute information to obtain body shape information of the target occupant.
Optionally, the determining, based on the target image, attribute information of the target occupant includes:
And determining attribute information of the target passenger based on the target image through an attribute identification model.
Optionally, the attribute information includes at least one of age and sex of the target occupant.
Optionally, the determining, based on the target image, three-dimensional feature information of a key point on the target occupant includes:
determining depth characteristic information of key points on the target passenger based on the target image through a depth identification model;
determining two-dimensional position characteristic information of the key points based on the target image through a position identification model;
and determining three-dimensional characteristic information of the key points based on the depth characteristic information and the two-dimensional position characteristic information.
Optionally, the depth characteristic information is used to indicate an absolute depth of the key point.
Optionally, the posture information includes at least one of posture information and sitting posture information of the target occupant.
Optionally, the adjusting the seat on which the target passenger sits based on the posture information of the target passenger includes:
determining target arrangement information of a seat on which the target occupant sits based on the posture information of the target occupant;
Determining adjustment information based on the current configuration information of the seat on which the target occupant sits and the target configuration information;
And adjusting the seat on which the target passenger sits based on the adjustment information.
Optionally, the camera is a monocular camera.
According to a second aspect of the present application, there is provided a seat adjusting device comprising:
The acquisition module is used for acquiring a target image of a target passenger in the target vehicle through a camera in the target vehicle;
the determining module is used for determining three-dimensional characteristic information of key points on the target passenger based on the target image and predicting body posture information of the target passenger based on the three-dimensional characteristic information of the key points;
and the adjusting module is used for adjusting the seat on which the target passenger sits based on the body posture information of the target passenger.
According to a third aspect of the present application there is also provided a computer readable storage medium having stored thereon a computer program which when executed by a processor performs steps in an embodiment of the present application.
According to a fourth aspect of the present application there is also provided a controller having stored thereon a computer program which when executed by a processor carries out the steps of an embodiment of the present application.
According to a fifth aspect of the present application, there is also provided a vehicle comprising a controller in an embodiment of the present application.
According to a sixth aspect of the present application there is also provided a computer program product comprising a computer program or instructions which when executed by a processor carries out the steps of an embodiment of the present application.
In summary, in the embodiment of the application, the target image of the target passenger in the target vehicle is acquired through the camera in the target vehicle, so that the cost can be reduced, the three-dimensional characteristic information of the key points on the target passenger is determined based on the target image, the body posture information of the target passenger is predicted based on the three-dimensional characteristic information of the key points, and the seat on which the target passenger sits is adjusted based on the body posture information of the target passenger, so that the accuracy of seat adjustment can be improved, and the application can reduce the cost, improve the accuracy of seat adjustment and further improve the sitting experience of the user.
Additional features and advantages of the application will be set forth in the detailed description which follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the description of the embodiments will be briefly described below. It is evident that the drawings in the following description are only some embodiments of the application and that other drawings may be obtained from these drawings without inventive effort for a person skilled in the art.
For a more complete understanding of the present application and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which like reference numerals represent like parts throughout the following description.
FIG. 1 is a flow chart of steps of a seat adjustment method provided in an exemplary embodiment of the present application;
FIG. 2 is a schematic illustration of another seat adjustment method provided in an exemplary embodiment of the present application;
FIG. 3 is a schematic illustration of another seat adjustment method provided in an exemplary embodiment of the present application;
FIG. 4 is a schematic illustration of another seat adjustment method provided in an exemplary embodiment of the present application;
Fig. 5 is a schematic view of a seat adjusting device provided in an exemplary embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by a person skilled in the art without any inventive effort, are intended to be within the scope of the present application based on the embodiments of the present application.
Referring to fig. 1, the seat adjusting method according to the embodiment of the application includes steps 100-300, which are described in detail below.
Step 100, acquiring a target image of a target passenger in a target vehicle through a camera in the target vehicle.
The camera comprises at least one camera, when the camera is one, the camera is a monocular camera, and the monocular camera refers to a single camera in a target vehicle. When the camera is a monocular camera, the cost can be further reduced.
The target occupant may be any occupant riding in the target vehicle, for example, the target occupant may be a driver or a passenger of the target vehicle. The target image of the target occupant may include at least upper body information of the target occupant.
It will be appreciated that a monocular camera may be used to capture the target image of a target occupant, or a monocular camera may capture the target image of multiple target occupants, as the embodiment is not limited in this respect.
According to the method and the device for adjusting the seat, the target image of the target passenger in the target vehicle is acquired through the monocular camera in the target vehicle, the related information of the target passenger does not need to be acquired through the camera outside the target vehicle, and the related information of the target passenger does not need to be acquired through the cameras, so that the cost consumed by adjusting the seat can be reduced.
Alternatively, the target vehicle may acquire the target image of the target occupant in the target vehicle through a monocular camera in the target vehicle when receiving the adjustment instruction, or the target vehicle may acquire the target image of the target occupant in the seat in response to detecting that the seat is seated, which is not limited herein.
Step 200, determining three-dimensional characteristic information of key points on the target passenger based on the target image, and predicting body posture information of the target passenger based on the three-dimensional characteristic information of the key points.
The key points refer to specific position points on the target passenger, the key points can comprise key points of an upper body of the target passenger, or the key points can comprise key points of the upper body and key points of a lower body of the target passenger, at the moment, the key points of the upper body of the target passenger can be determined firstly based on the target image, and then the key points of the lower body of the target passenger can be determined based on the key points of the upper body.
When the key point is the key point of the upper body of the target occupant, the key point may be, for example, at least one of a key point on the eyes, a key point on the nose, a key point on the ears, a key point on the shoulders, a key point of the hands, a key point of the thighs, a key point of the knees, and a key point of the trunk.
The three-dimensional feature information of the key points may refer to three-dimensional position information of the key points. The posture information of the target occupant may include at least one of posture information and sitting posture information of the target occupant.
The body type information of the target occupant is used to describe the external form and outline of the target occupant, and may be set according to actual conditions, for example, the body type information of the target occupant may include at least one of height information, arm length information, and leg length information of the target occupant.
The sitting posture information refers to the position and posture of each part of the body when the target occupant sits.
In this embodiment, the target image is identified to obtain three-dimensional feature information of a key point on the target occupant, and body posture information of the target occupant is predicted based on the three-dimensional feature information of the key point, so that the seat is adjusted based on the body posture information of the target occupant, and thus accuracy of seat adjustment is improved.
In some embodiments, the manner of determining the three-dimensional feature information of the key points on the target occupant may be set according to the actual situation, for example, the three-dimensional feature information of the key points on the target occupant may be determined by using a directional gradient histogram (Histogram of Oriented Gradients, HOG) or a trained neural network model based on the target image, which is not limited herein.
When determining three-dimensional feature information of a key point on a target occupant based on a target image by a trained neural network model, determining the three-dimensional feature information of the key point on the target occupant based on the target image includes:
Determining depth characteristic information of key points on a target passenger based on the target image through a depth identification model;
determining two-dimensional position characteristic information of the key points based on the target image through a position identification model;
and determining three-dimensional characteristic information of the key points based on the depth characteristic information and the two-dimensional position characteristic information.
The depth feature information is used for indicating the depth of the key point, namely the depth feature information is used for representing the third-dimensional position feature information of the key point, so that the depth feature information and the two-dimensional position feature information form the three-dimensional feature information of the key point.
The depth recognition model and the position recognition model are both trained neural network models, and the types of the depth recognition model and the position recognition model can be set according to actual situations, for example, the depth recognition model and the position recognition model can be YOLOv7-pose models or convolutional neural network models, and the embodiment is not limited herein.
In the embodiment, depth characteristic information of key points on a target passenger is determined based on a target image through a depth recognition model, two-dimensional position characteristic information of the key points is determined based on the target image through a position recognition model, three-dimensional characteristic information of the key points is determined based on the depth characteristic information and the two-dimensional position characteristic information, three-dimensional characteristic information is obtained through a trained neural network model, and accuracy of the three-dimensional characteristic information is improved, so that accuracy of adjusting a seat based on the three-dimensional characteristic information is improved.
It will be appreciated that the depth recognition model and the position recognition model need to be trained before determining depth characteristic information of a key point on a target occupant based on a target image by the depth recognition model and determining two-dimensional position characteristic information of the key point based on the target image by the position recognition model. The training process for the depth recognition model and the position recognition model can be as follows:
acquiring a first training sample set, wherein the first training sample set comprises at least one first sample image;
determining sample depth feature information of sample key points based on the first sample image through an initial depth recognition model;
determining sample two-dimensional position feature information of sample key points based on the first sample image through an initial position identification model;
obtaining sample three-dimensional characteristic information of sample key points based on the sample depth characteristic information and the sample two-dimensional position characteristic information;
Determining a first loss value based on the sample three-dimensional characteristic information and label information corresponding to the first sample image;
Based on the first loss value, training the initial depth recognition model and the position recognition model to obtain the depth recognition model and the position recognition model.
The first training sample set may be a sample set subjected to a sample enhancement process, which may be, for example, at least one of rotation, scaling, and translation.
The first sample image may include at least one of an image obtained by photographing a sample occupant in a different vehicle, an image obtained by photographing a sample occupant in a vehicle of a different environment, and an image obtained by photographing a sample occupant at a different viewing angle.
The first loss value may be a value of the first loss function, or the first loss value may include a value of the first loss function and a first regularization value. The type of the first loss function may be set according to practical situations, for example, the first loss function may be a mean square error (Mean Squared Error, MSE) or a cross entropy function, which is not limited herein.
The manner of determining the first regularization value may be set according to practical situations, for example, the first regularization value may be determined by Dropout, L1 or L2, which is not limited herein.
The overfitting phenomenon may be alleviated when the loss value comprises a value of the first loss function and a regularized value.
In some embodiments, a learning rate (LEARNING RATE, LR) decay mechanism may be employed during training of the initial depth recognition model and the location recognition model, with progressive decreases in learning rate as training progresses to improve the stability and generalization ability of the model.
In some embodiments, after the depth recognition model and the location recognition model are obtained, a model quantization process and/or pruning process may be performed on the depth recognition model and the location recognition model to reduce a storage space and an inference space of the model.
In some embodiments, depth characteristic information is used to indicate the absolute depth of the keypoints. In this embodiment, the depth feature information is used to indicate the absolute depth of the key point, so that the body posture information obtained based on the three-dimensional feature information of the key point is more accurate, and the adjustment of the seat based on the body posture information is more accurate.
In some embodiments, the manner of predicting the posture information of the target occupant based on the three-dimensional feature information of the key points may be set according to the actual situation, and the embodiment is not limited herein.
For example, the body posture information of the target passenger can be predicted based on the three-dimensional characteristic information of the key points through a preset body posture prediction model.
For another example, the posture information includes body shape information, and predicting the body shape information of the target occupant based on the three-dimensional feature information of the key points includes:
Predicting volume information of the target passenger based on the three-dimensional characteristic information of the key points;
based on the volume information, body type information of the target occupant is predicted.
The target vehicle can be modeled based on the three-dimensional characteristic information of the key points, so that the volume information of the target passenger is obtained, and then the body type information of the target passenger is predicted based on the volume information.
Specifically, the body type information of the target occupant may be predicted based on the volume information by a predictive mapping strategy. The type of the preset mapping policy may be set according to actual situations, for example, the preset mapping policy may be a preset mapping table or a preset function, which is not limited herein.
In this embodiment, based on the three-dimensional feature information of the key points, the volume information of the target occupant is predicted, and based on the volume information, the body type information of the target occupant is predicted, so that the accuracy of the body type information is improved, and the accuracy of adjusting the seat based on the body type information is further improved.
In some embodiments, predicting body type information of the target occupant based on the volume information includes:
Predicting weight information of the target occupant based on the volume information;
based on the weight information and the three-dimensional feature information of the key points, body shape information of the target occupant is predicted.
Wherein the weight information of the target occupant can be predicted based on the volume information and the density information. The target vehicle predicts initial body type information of the target passenger based on the three-dimensional characteristic information of the key points, and then adjusts the initial body type information based on the weight information to obtain body type information of the target passenger.
In this embodiment, the weight information of the target occupant is predicted based on the volume information, and the body type information of the target occupant is predicted based on the weight information and the three-dimensional feature information of the key points, so that the accuracy of the body type information is further improved, and the accuracy of adjusting the seat based on the body type information is further improved.
In some embodiments, the manner of predicting the initial body shape information of the target occupant based on the three-dimensional feature information of the key points may be selected according to practical situations, for example, the initial body shape information of the target occupant may be predicted based on the three-dimensional feature information of the key points by a human engineering method, or the initial body shape information of the target occupant may be predicted based on the three-dimensional feature information of the key points by a body shape neural network model, which is not limited herein.
When predicting initial body shape information of the target occupant based on the three-dimensional feature information of the key points by the ergonomic method, the process of predicting body shape information of the target occupant based on the body weight information and the three-dimensional feature information of the key points may be:
Predicting initial body type information of a target passenger based on three-dimensional characteristic information of key points by a human engineering method;
Based on the weight information, the initial body shape information is subjected to adjustment processing to obtain body shape information of the target occupant.
Among them, ergonomic methods, also called ergonomics or ergonomics, are techniques that study interactions between a person and other elements, such as tools, devices, machines or environments, etc.
The method for adjusting the initial body type information based on the body weight information may be set according to actual situations, for example, by presetting a first mapping policy, the body weight information and the initial body type information are mapped, so as to obtain body type information.
The type of the preset first mapping policy may be selected according to practical situations, for example, the preset first mapping policy may be a first mapping table or a function, which is not limited in this embodiment.
In this embodiment, the three-dimensional feature information of the key points is used to predict the initial body shape information of the target occupant, and the body shape information of the target occupant is obtained by adjusting the initial body shape information based on the weight information.
In some embodiments, the key points are key points of the upper body of the target occupant, the initial body shape information includes initial height information, and predicting the initial body shape information of the target occupant based on three-dimensional feature information of the key points by an ergonomic method includes:
Predicting the height information of the upper body of the target passenger based on the three-dimensional characteristic information of the key points by an artificial engineering method;
Predicting the height information of the lower body of the target occupant based on the height information of the upper body;
based on the height information of the upper body and the height information of the lower body, initial height information of the target occupant is determined.
The three-dimensional feature information of the key points of the upper body may be determined based on the three-dimensional feature information of the key points of the upper body, the height information of the upper body of the target occupant may be predicted based on the three-dimensional feature information of the key points of the upper body, and the height information of the lower body of the target occupant may be predicted based on the three-dimensional feature information of the key points of the lower body. The height information of the lower body of the target occupant may also be referred to as leg length information of the target occupant.
In this embodiment, the height information of the upper body of the target occupant is predicted based on the three-dimensional feature information of the key points by the ergonomic method, the height information of the lower body of the target occupant is predicted based on the height information of the upper body, and the initial height information of the target occupant is determined based on the height information of the upper body and the height information of the lower body, so that the height information of the lower body which is not shot is determined by the ergonomic method, the human feature information of the target occupant is enriched, the accuracy of the height information is further improved, and the accuracy of adjusting the seat based on the height information is further improved.
In some embodiments, the present embodiment further comprises:
determining attribute information of a target passenger based on the target image;
Based on the weight information, the initial body type information is adjusted to obtain body type information of the target passenger, including:
and adjusting the initial body type information based on the weight information and the attribute information to obtain the body type information of the target passenger.
The attribute information of the target occupant may refer to data describing the demographics of the target occupant, which may be set according to actual situations, for example, the attribute information of the target occupant may include at least one of age and sex of the target occupant, which is not limited herein.
The method for adjusting the initial body type information based on the body weight information and the attribute information may be set according to actual situations, for example, by presetting a second mapping policy, the body weight information, the attribute information and the initial body type information are mapped, so as to obtain body type information.
The type of the preset second mapping policy may be selected according to practical situations, for example, the preset second mapping policy may be a second mapping table or a function, which is not limited herein.
In this embodiment, based on the target image, the attribute information of the target occupant is determined, and based on the weight information and the attribute information, the initial body type information is adjusted to obtain the body type information of the target occupant, so that the body type information is obtained based on the weight information and the attribute information, and the accuracy of the body type information is further improved, so that the accuracy of adjusting the seat based on the body type information is further improved.
In some embodiments, determining attribute information of the target occupant based on the target image includes:
And determining attribute information of the target passenger based on the target image through the attribute identification model.
The attribute identifying model may be a trained neural network model, and the type of the model may be set according to practical situations, for example, the attribute identifying model may be MobieNet or a convolutional neural network model, which is not limited herein.
In the embodiment, the attribute information of the target passenger is determined based on the target image through the attribute identification model, so that the trained neural network model is realized, the attribute information of the target passenger is obtained, and the accuracy of the attribute information is improved.
It may be appreciated that, before determining, based on the target image, the attribute information of the target occupant through the attribute recognition model, the attribute recognition model may be trained, and the process of training the attribute recognition model may be:
Acquiring a second training sample set, wherein the second training sample set comprises at least one second sample image;
determining sample attribute information of the sample occupant based on the second sample image through the initial attribute identification model;
determining a second loss value based on the sample attribute information and label information corresponding to the second sample image;
And training the initial attribute identification model based on the second loss value to obtain an attribute identification model.
Wherein the second training sample set may be a sample set subjected to a sample enhancement process, which may be, for example, at least one of rotation, scaling and translation. The second training sample set and the first training sample set may or may not be the same sample set.
The second sample image may include at least one of an image photographed by a different sample occupant, an image photographed by a sample occupant of a different environment, and an image photographed by a sample occupant at a different viewing angle.
When the attribute information includes sex, the tag information of the second sample image may include "0" or "1", with "0" indicating male and "1" indicating female. When the attribute information includes age, the tag information of the second sample image may include "18-25 years old", "26-30 years old", "31-35 years old", "36-40 years old", "41-45 years old", "46-50 years old", "51-55 years old", or "56-60 years old".
The second loss value may be a value of the loss function, or the second loss value may include a value of the second loss function and a second regularized value. The type of the second loss function may be set according to practical situations, for example, the loss function may be a mean square error (Mean Squared Error, MSE) or a cross entropy function, which is not limited herein.
The manner of determining the second regularization value may be set according to practical situations, for example, the regularization value may be determined by Dropout, L1 or L2, which is not limited herein.
The overfitting phenomenon may be alleviated when the loss value comprises a value of the second loss function and a regularized value.
In some embodiments, a learning rate (LEARNING RATE, LR) decay mechanism may be employed during training of the initial attribute identification model, with progressively decreasing learning rates as training progresses to improve the stability and generalization ability of the model.
In some embodiments, after the attribute identification model is obtained, a model quantization process and/or pruning process may be performed on the attribute identification model to reduce the storage space and inference space of the model.
In some embodiments, when the posture information includes sitting posture information, the sitting posture information may be directly determined based on the three-dimensional feature information of the key points, or the posture information may be determined based on the three-dimensional feature information of the key points, and then the sitting posture information may be determined based on the posture information.
Step 300, adjusting the seat on which the target passenger sits based on the posture information of the target passenger.
Wherein, adjusting the seat on which the target occupant sits may refer to adjusting at least one of the position, the angle, the degree of bulging of the lumbar support portion of the seat, and the position of the headrest on the seat.
In some embodiments, based on the body posture information of the target occupant, the process of adjusting the seat on which the target occupant sits may be:
Determining target configuration information of a seat on which the target occupant sits based on body posture information of the target occupant;
determining adjustment information based on current configuration information and target configuration information of a seat on which a target occupant sits;
based on the adjustment information, the seat on which the target occupant sits is adjusted.
The third mapping strategy can be preset to map the body posture information of the target passenger, so that the target configuration information of the seat on which the target passenger sits is obtained. The type of the preset third mapping policy may be set according to actual situations, for example, the preset third mapping policy may be a third mapping table or a function, which is not limited in this embodiment.
The third mapping table or function is constructed based on configuration information of seats liked by the user of the different posture information. The current configuration information of the seat refers to at least one of current position information of the seat, current angle information, a current degree of protrusion of a lumbar support portion of the seat, and a current position of a headrest on the seat. The target configuration information of the seat refers to at least one of target position information of the seat, target angle information, target bulging degree of a lumbar support portion of the seat, and target position of a headrest on the seat.
The adjusting information is used for indicating the adjusting quantity of the seat, and after the seat is adjusted based on the adjusting information, the target passenger can be in a comfortable state, so that the riding comfort of the target passenger is improved, and the riding experience of the target passenger is improved.
In some embodiments, the target vehicle may include an RK3568 chip, a camera head is coupled through the RK3568 chip, and the seat adjustment method provided by the present application is performed through the RK3568 chip.
The seat adjusting method provided by the application is further described below with reference to fig. 2.
The method comprises the steps of acquiring a target image of a target passenger in a target vehicle through a monocular camera in the target vehicle, determining three-dimensional characteristic information of key points on the target passenger through a depth recognition model and a position recognition model through a computing center in the target vehicle, determining attribute information of the target passenger through an attribute recognition model based on the target image, uploading the three-dimensional characteristic information of the key points, the attribute information and current configuration information of a seat to a cloud computing center, determining adjusting information through the cloud computing center based on the three-dimensional characteristic information and the current configuration information of the seat, returning the adjusting information to the computing center in the target vehicle, sending the adjusting information to a control center through the computing center in the target vehicle, adjusting the seat taken by the target passenger through the control center based on the adjusting information, and displaying the adjusted seat on a display device in the target vehicle.
The specific implementation manner and the corresponding beneficial effects of this embodiment may refer to the above-mentioned seat adjusting method embodiment, and this embodiment is not described herein again.
The seat adjusting method provided by the application is further described below with reference to fig. 3. In this embodiment, the seat adjustment method includes a training process of a model, an reasoning process of the model, and a determination process.
The training process of the model comprises the following steps:
Acquiring a first sample image and a second sample image through a monocular camera in a vehicle, labeling a depth label on the first sample image, labeling a two-dimensional position label and an attribute label on the second sample image, preprocessing the first sample image, preprocessing the second sample image, constructing an initial depth recognition model, constructing an initial position recognition model and an initial attribute recognition model, training the initial depth recognition model through the first sample image to obtain a depth recognition model, and training the initial position recognition model and the initial attribute recognition model through the second sample image to obtain a position recognition model and an attribute recognition model.
The reasoning process of the model comprises:
the method comprises the steps of acquiring a target image of a target passenger in a target vehicle through a monocular camera in the target vehicle, preprocessing the target image to obtain a processed image, obtaining depth information of a key point of the upper body through a depth recognition model based on the processed image, obtaining two-dimensional position characteristic information of the key point of the upper body through a position recognition model based on the processed image, determining three-dimensional characteristic information of the key point of the upper body based on the depth characteristic information and the two-dimensional position characteristic information, and obtaining attribute information of the target passenger based on the target image through an attribute recognition model.
The determining process comprises the following steps:
The body weight information is determined based on the three-dimensional characteristic information of the upper body key points, the lower body information is determined based on the three-dimensional characteristic information of the upper body key points by a human engineering method, the initial body shape information of the target passenger is determined based on the upper body key points and the lower body information, the initial body shape information is adjusted based on the body weight information and the attribute information, the body shape information is obtained, the sitting posture information is determined based on the body shape information, and the seat is adjusted based on the sitting posture information.
The specific implementation manner and the corresponding beneficial effects of this embodiment may refer to the above-mentioned seat adjusting method embodiment, and this embodiment is not described herein again.
The seat adjusting method provided by the application is further described below with reference to fig. 4. In the present embodiment, a driver is taken as a target passenger as an example.
The driver takes the driving seat of the target vehicle, the target vehicle is powered on, and the monocular camera in the target vehicle is powered on. And acquiring a target image of the driver through a monocular camera. The method comprises the steps that a computing center in a target vehicle determines weight information and attribute information of a driver based on a target image, initial body type information of the target passenger is determined through a human engineering method, body type information is determined based on the weight information, the attribute information and the initial body type information through the computing center in the target vehicle, sitting posture information is determined based on the body type information, adjustment information of a driver seat is determined based on the sitting posture information and current configuration information of the driver seat, the adjustment information is sent to a control center, and the control center adjusts the driver seat to a corresponding position based on the adjustment information.
And if the manual adjustment instruction of the driver is received, adjusting the driving seat based on the manual adjustment instruction.
The specific implementation manner and the corresponding beneficial effects of this embodiment may refer to the above-mentioned seat adjusting method embodiment, and this embodiment is not described herein again.
From the above, in the embodiment of the application, the target image of the target passenger in the target vehicle is acquired through the camera in the target vehicle, so that the cost can be reduced, the three-dimensional characteristic information of the key points on the target passenger is determined based on the target image, the body posture information of the target passenger is predicted based on the three-dimensional characteristic information of the key points, and the seat on which the target passenger sits is adjusted based on the body posture information of the target passenger, so that the accuracy of seat adjustment can be improved, and the application can improve the accuracy of seat adjustment while reducing the cost and further improve the sitting experience of the user.
Fig. 5 is a schematic structural view of a seat adjusting device according to an embodiment of the present application. Referring to fig. 5, the seat adjusting device may include an acquisition module 501, a determination module 502, and an adjustment module 503. Wherein:
An acquiring module 501 is configured to acquire, by using a camera in a target vehicle, a target image of a target occupant in the target vehicle.
The determining module 502 is configured to determine three-dimensional feature information of a key point on the target occupant based on the target image, and predict body posture information of the target occupant based on the three-dimensional feature information of the key point.
An adjusting module 503, configured to adjust a seat on which the target occupant sits based on body posture information of the target occupant.
The acquiring module 501, the determining module 502 and the adjusting module 503 may be respectively configured to perform all the steps in the embodiments corresponding to the seat adjusting method, and specific implementation manners and more details of these modules may refer to corresponding method portions, which are not described herein in detail.
Embodiments of the present application also provide a computer-readable storage medium having instructions stored thereon that, when executed by a processor, cause the processor to be configured to perform the seat adjustment method described above.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.
Computer-readable media include both permanent and non-permanent, removable and non-removable media, and information storage may be implemented by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated communication signals and carrier waves.
In some embodiments, the present application also provides a controller having a computer program stored thereon, which when executed by a processor, implements the steps of embodiments of the present application. The controller may be located in the target vehicle or in the server. When the controller is located in other servers, after the controller obtains the body posture information or the adjustment information, the body posture information or the adjustment information is sent to the target vehicle, and the target vehicle adjusts the seat based on the body posture information or the adjustment information.
In some embodiments, the application further provides a vehicle comprising the controller according to the embodiments of the application. In this embodiment, the vehicle may be a fuel-oil vehicle, a plug-in hybrid vehicle, a new energy vehicle, or the like, which is not particularly limited in this embodiment.
In some embodiments, the application also provides a computer program product comprising a computer program or instructions which, when executed by a processor, implement the steps in embodiments of the application.
In the description of the present application, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more features. In the description of the present application, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.
The embodiments, the implementation modes and the related technical features of the application can be mutually combined and replaced under the condition of no conflict.
The foregoing is only a preferred embodiment of the present application, and is not intended to limit the present application in any way, but any simple modification, equivalent variation and modification made to the above embodiment according to the technical matter of the present application still fall within the scope of the technical solution of the present application.
Claims (18)
1. A seat adjustment method, comprising:
acquiring a target image of a target passenger in a target vehicle through a camera in the target vehicle;
Determining three-dimensional characteristic information of key points on the target passenger based on the target image, and predicting body posture information of the target passenger based on the three-dimensional characteristic information of the key points;
And adjusting the seat on which the target passenger sits based on the posture information of the target passenger.
2. The method according to claim 1, wherein the posture information includes body shape information, and the predicting the body shape information of the target occupant based on the three-dimensional feature information of the key points includes:
predicting volume information of the target passenger based on the three-dimensional feature information of the key points;
based on the volume information, body type information of the target occupant is predicted.
3. The method according to claim 2, wherein predicting body type information of the target occupant based on the volume information includes:
Predicting weight information of the target occupant based on the volume information;
And predicting body type information of the target passenger based on the weight information and the three-dimensional feature information of the key points.
4. The method according to claim 3, wherein predicting body type information of the target occupant based on the body weight information and the three-dimensional feature information of the key points includes:
Predicting initial body type information of the target passenger based on the three-dimensional characteristic information of the key points by a human engineering method;
and based on the weight information, performing adjustment processing on the initial body type information to obtain body type information of the target passenger.
5. The method of claim 4, wherein the keypoints are keypoints of an upper body of the target occupant, the initial body shape information comprises initial height information, and the predicting the initial body shape information of the target occupant based on three-dimensional feature information of the keypoints by an ergonomic method comprises:
predicting the height information of the upper body of the target passenger based on the three-dimensional characteristic information of the key points by an artificial engineering method;
predicting height information of a lower body of the target occupant based on the height information of the upper body;
Initial height information of the target occupant is determined based on the height information of the upper body and the height information of the lower body.
6. The method as recited in claim 4, further comprising:
determining attribute information of the target occupant based on the target image;
the step of performing adjustment processing on the initial body type information based on the weight information to obtain body type information of the target passenger includes:
and adjusting the initial body type information based on the weight information and the attribute information to obtain the body type information of the target passenger.
7. The method of claim 6, wherein the determining attribute information of the target occupant based on the target image comprises:
and determining attribute information of the target passenger based on the target image through an attribute identification model.
8. The method of claim 6, wherein the attribute information includes at least one of an age and a gender of the target occupant.
9. The method of claim 1, wherein the determining three-dimensional feature information of a keypoint on the target occupant based on the target image comprises:
Determining depth characteristic information of key points on the target passenger based on the target image through a depth recognition model;
determining two-dimensional position characteristic information of the key points based on the target image through a position identification model;
And determining three-dimensional characteristic information of the key points based on the depth characteristic information and the two-dimensional position characteristic information.
10. The method of claim 9, wherein the depth characteristic information is used to indicate an absolute depth of the keypoint.
11. The method of claim 1, wherein the posture information includes at least one of body type information and sitting posture information of the target occupant.
12. The method of claim 1, wherein adjusting the seat on which the target occupant sits based on the body posture information of the target occupant comprises:
determining target configuration information of a seat on which the target occupant sits based on body posture information of the target occupant;
Determining adjustment information based on current configuration information of a seat on which the target occupant sits and the target configuration information;
Based on the adjustment information, the seat on which the target occupant sits is adjusted.
13. The method of any one of claims 1-12, wherein the camera is a monocular camera.
14. A seat adjustment apparatus, comprising:
The acquisition module is used for acquiring a target image of a target passenger in the target vehicle through a camera in the target vehicle;
the determining module is used for determining three-dimensional characteristic information of key points on the target passenger based on the target image and predicting body posture information of the target passenger based on the three-dimensional characteristic information of the key points;
and the adjusting module is used for adjusting the seat on which the target passenger sits based on the body posture information of the target passenger.
15. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the method according to any of claims 1 to 13.
16. A controller having a computer program stored thereon, which when executed by a processor performs the steps of the method according to any of claims 1 to 13.
17. A vehicle is characterized in that, comprising a controller according to claim 16.
18. A computer program product comprising a computer program or instructions which, when executed by a processor, implement the steps of the method of any one of claims 1 to 13.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510138454.3A CN119872358A (en) | 2025-02-07 | 2025-02-07 | Seat adjusting method, device, storage medium, controller, vehicle and product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510138454.3A CN119872358A (en) | 2025-02-07 | 2025-02-07 | Seat adjusting method, device, storage medium, controller, vehicle and product |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119872358A true CN119872358A (en) | 2025-04-25 |
Family
ID=95438951
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510138454.3A Pending CN119872358A (en) | 2025-02-07 | 2025-02-07 | Seat adjusting method, device, storage medium, controller, vehicle and product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119872358A (en) |
-
2025
- 2025-02-07 CN CN202510138454.3A patent/CN119872358A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11983850B2 (en) | Image processing method and apparatus, device, and storage medium | |
| US11960533B2 (en) | Visual search using multi-view interactive digital media representations | |
| CN110399825B (en) | Facial expression migration method and device, storage medium and computer equipment | |
| EP4404148A1 (en) | Image processing method and apparatus, and computer-readable storage medium | |
| CN111071187A (en) | Driving environment intelligent adjustment and driver registration method and device, vehicle and equipment | |
| CN103177269B (en) | For estimating the apparatus and method of object gesture | |
| CN111226258B (en) | Signal conversion system and signal conversion method | |
| CN111881886A (en) | Intelligent seat control method and device based on posture recognition | |
| US12217395B2 (en) | Exemplar-based object appearance transfer driven by correspondence | |
| CN110148179A (en) | A kind of training is used to estimate the neural net model method, device and medium of image parallactic figure | |
| US20190114462A1 (en) | Signal translation system and signal translation method | |
| CN119295638B (en) | Image rendering method, device, equipment, storage medium and program product based on Gaussian splashing | |
| CN112733672A (en) | Monocular camera-based three-dimensional target detection method and device and computer equipment | |
| CN109920512B (en) | Training method and device for three-dimensional dose distribution network model | |
| CN114926591B (en) | Multi-branch deep learning 3D face reconstruction model training method, system and medium | |
| CN113386638A (en) | Method and device for adjusting vehicle seat | |
| CN119329548B (en) | Vehicle control method, apparatus, device, readable storage medium, and program product | |
| CN112613445A (en) | Face image generation method and device, computer equipment and storage medium | |
| CN119872358A (en) | Seat adjusting method, device, storage medium, controller, vehicle and product | |
| CN113044045B (en) | Self-adaptive adjustment method for seats in intelligent cabin | |
| CN113901884B (en) | Face pose estimation method, system, device and storage medium based on feature matching | |
| CN113469091B (en) | Face recognition method, training method, electronic device and storage medium | |
| CN114078149B (en) | Image estimation method, electronic device and storage medium | |
| CN110610184A (en) | Method, device and equipment for detecting salient object of image | |
| CN113496472A (en) | Image defogging model construction method, road image defogging device and vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |