CN111899186A - Image data enhancement method, system, storage medium and storage yard coverage detection method - Google Patents
Image data enhancement method, system, storage medium and storage yard coverage detection method Download PDFInfo
- Publication number
- CN111899186A CN111899186A CN202010601516.7A CN202010601516A CN111899186A CN 111899186 A CN111899186 A CN 111899186A CN 202010601516 A CN202010601516 A CN 202010601516A CN 111899186 A CN111899186 A CN 111899186A
- Authority
- CN
- China
- Prior art keywords
- image data
- data enhancement
- parameters
- enhancement
- target
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides an image data enhancement method, a system, a storage medium and a storage yard coverage detection method, wherein the image data enhancement method comprises the following steps: acquiring a data enhancement operation instruction, wherein the data enhancement operation instruction comprises data enhancement operation and parameters of the data enhancement operation; training parameters of the data enhancement operation, and determining target parameters of the data enhancement operation; and enhancing the image data according to the data enhancement operation and the target parameter. Compared with the prior art, the image data enhanced by the image data enhancement method provided by the invention is very close to the original data, the quantity of the image data is effectively increased, the enhancement effect is better, and the diversity and distribution similarity of the enhanced image data are ensured.
Description
Technical Field
The invention relates to the field of data enhancement, in particular to an image data enhancement method, an image data enhancement system, a computer-readable storage medium and a storage yard coverage detection method.
Background
The existing image data enhancement method is a digital image processing method, such as rotation, translation, clipping and color transformation, but parameters such as the rotation angle, the translation pixel value and the like need to be set by human experience or randomly, uncertainty exists, different parameters have different effects on image data enhancement, and the final detection effect can be influenced. In addition, if the parameters of the enhancement operation are not considered, the diversity of the data is indeed enhanced, but the distribution similarity of the data is destroyed, so an image data enhancement method is needed to maximize the enhancement effect and ensure the diversity and distribution similarity of the enhanced image data.
Disclosure of Invention
The present invention is directed to solving at least one of the problems of the prior art or the related art.
To this end, a first aspect of the invention proposes an image data enhancement method.
A second aspect of the invention is directed to an image data enhancement system.
A third aspect of the invention is directed to a computer-readable storage medium.
The fourth aspect of the present invention is to provide a yard coverage detection method.
In view of the above, according to a first aspect of the present invention, there is provided an image data enhancement method, including: acquiring a data enhancement operation instruction, wherein the data enhancement operation instruction comprises data enhancement operation and parameters of the data enhancement operation; training parameters of the data enhancement operation, and determining target parameters of the data enhancement operation; and enhancing the image data according to the data enhancement operation and the target parameter.
According to the image data enhancement method provided by the invention, the target parameters of data enhancement are determined by training the parameters of the data enhancement operation, and then the image data is enhanced according to the data enhancement operation and the target parameters. Compared with the prior art, the image data enhanced by the image data enhancement method provided by the invention is very close to the original data, the quantity of the image data is effectively increased, the enhancement effect is better, and the diversity and distribution similarity of the enhanced image data are ensured.
In the above technical solution, further, training parameters of the data enhancement operation to determine target parameters of the data enhancement operation specifically includes: determining a discrete search space according to the data enhancement operation instruction; sampling in a discrete search space through a neural network controller to obtain a target operation instruction; enhancing the image data according to the target operation instruction to generate an image data sample; generating a sub-detection network according to the image data sample, and outputting the accuracy of the image data sample through the sub-detection network; and fitting the parameters in the target operation instruction through the neural network controller according to the accuracy of the image data sample, and determining the parameters in the target operation instruction as the target parameters when the fitting is achieved.
In the technical scheme, the process of determining the target parameters is abstracted into a process of searching for a discrete search optimal solution, and the learning result of the neural network controller is used as feedback to judge whether the parameters in the target operation instruction are the target parameters, so that the aim of enhancing the image data is fulfilled, the effect after enhancing the image data is better, meanwhile, the parameter of data enhancement operation corresponding to the whole image data set is prevented from being trained, and the efficiency of determining the target parameters is improved.
In the above technical solution, further, after the accuracy of the image data sample output by the sub-detection network, the method further includes: and preprocessing the accuracy of the image data sample through a reinforcement learning algorithm, and sending the accuracy of the preprocessed image data sample to a neural network controller.
In the technical scheme, the accuracy of the image data sample is subjected to quantitative processing through a reinforcement learning algorithm, and the accuracy of the image data sample subjected to quantitative processing is sent to a neural network controller, that is, the neural network controller fits parameters in a target operation instruction according to the accuracy information subjected to quantitative processing, and determines the parameters in the target operation instruction as target parameters. The efficiency of determining the target parameters is improved by the cooperation of the reinforcement learning algorithm and the neural network controller.
In the above technical solution, further, the image data enhancement method further includes: and fitting the parameters in the target operation instruction through the neural network controller, and sampling again in the discrete search space through the neural network controller when the fitting is not achieved.
According to the technical scheme, the parameters in the target operation instruction are fitted through the neural network controller, and when the fitting is not achieved, the neural network controller is used for sampling again in the discrete search space, namely, the neural network controller is used for sampling for multiple times in the discrete search space. The neural network controller adjusts the sampling direction according to the fitting result, and then iterative training of parameters in the target operation instruction is achieved, so that the efficiency of determining the target operation parameters is higher, the determined target parameters are optimal solutions or approximate optimal solutions infinitely close to the optimal solutions, and the requirement of data enhancement is met.
In any of the above technical solutions, further, the data enhancement operation includes: any one or more of a cropping operation, a geometric transformation operation, a rotation operation, an automatic contrast operation, an inversion operation, a color equalization operation, a color transformation operation, a luminance transformation operation, a sharpness transformation operation, and a histogram equalization operation.
In the technical scheme, the data enhancement operation comprises any one or more of ten operations, namely a clipping operation, a geometric transformation operation, a rotation operation, an automatic contrast operation, a reversal operation, a color equalization operation, a color transformation operation, a brightness transformation operation, a definition transformation operation and a histogram equalization operation.
In the above technical solution, further, when the data enhancement operation includes X data enhancement operationsDetermining a seed enhancement strategy, wherein each seed enhancement strategy comprises b data enhancement operations, a, b and X are positive integers; the discrete space size of any data enhancement operation is: ((N X P M)b)aWhere N is the number of operations of the data enhancement operation, P is the discrete value of the probability of use of any kind of data enhancement operation, and M is the discrete value of the strength of any kind of data enhancement operation.
In this solution, the size of the discrete space is determined according to the data enhancement operation.
In the above technical solution, further, the reinforcement learning algorithm is a PPO algorithm, and a formula of the PPO algorithm is:
wherein, θ and θ 'are sub-enhancement strategies, KL (θ, θ') is the KL divergence of the two sub-enhancement strategies, β is a penalty coefficient, and β is 1 e-4.
In the technical scheme, the reinforcement learning algorithm is a PPO algorithm, penalty compensation is performed according to a formula, the corresponding accuracy of the obtained image data sample is higher, and the target parameter is more accurate.
According to a second aspect of the present invention, there is provided an image data enhancement system, a memory storing a computer program, and a processor executing the computer program to perform the steps of the image data enhancement method of any one of the above. Therefore, the image data enhancement system comprises all the beneficial effects of the image data enhancement method of any one of the technical schemes.
According to a third aspect of the present invention, a computer-readable storage medium is proposed, on which a computer program is stored, which computer program, when being executed by a processor, performs the steps of the image data enhancement method of any of the above. Therefore, the computer readable storage medium includes all the advantages of the image data enhancement method of any one of the above technical solutions.
According to a fourth aspect of the present invention, there is provided a yard coverage detection method comprising the image data enhancement method of any one of the above, wherein the image data is acquired yard coverage image data. Therefore, the storage yard coverage detection method has all the beneficial effects of the image data enhancement method of any one of the technical schemes.
Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 shows a schematic flow diagram of an image data enhancement method of one embodiment of the present invention;
FIG. 2 is a flow chart illustrating an image data enhancement method according to another embodiment of the present invention;
FIG. 3 shows a schematic flow diagram of an image data enhancement method according to a further embodiment of the invention;
fig. 4 shows a flow chart of an image data enhancement method according to another embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings. It should be noted that the embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those specifically described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
An embodiment of the first aspect of the present invention provides an image data enhancement method, and fig. 1 shows a flowchart of the image data enhancement method according to an embodiment of the present invention. Wherein, the method comprises the following steps:
102, acquiring a data enhancement operation instruction, wherein the data enhancement operation instruction comprises data enhancement operation and parameters of the data enhancement operation;
104, training parameters of the data enhancement operation, and determining target parameters of the data enhancement operation;
and 106, enhancing the image data according to the data enhancement operation and the target parameters.
In the image data enhancement method provided by this embodiment, the target parameter of data enhancement is determined by training the parameter of the data enhancement operation, and then the image data is enhanced according to the data enhancement operation and the target parameter. Compared with the prior art, the image data enhanced by the image data enhancement method provided by the invention is very close to the original data, the quantity of the image data is effectively increased, the enhancement effect is better, and the diversity and distribution similarity of the enhanced image data are ensured.
Fig. 2 shows a flow chart of an image data enhancement method according to still another embodiment of the present invention, as shown in fig. 2, the method includes:
In this embodiment, the process of determining the target parameters is abstracted to a process of finding a discrete search optimal solution, first, determining discrete search space according to the data enhancement operation instruction, sampling by a neural network controller to obtain a target operation instruction, wherein the target operation instruction comprises parameters of data enhancement operation and data enhancement operation, and generates an image data sample according to the target operation instruction, secondly, the accuracy of the image data sample output by the sub-detection network is input into the neural network controller as a reward signal, and finally, the neural network controller continuously learns, and fitting the parameters in the target operation instruction, wherein when the parameters are fitted, the parameters in the target operation instruction are the parameters of the optimal data enhancement operation or the parameters close to the optimal data enhancement operation, namely the target parameters. The learning result of the neural network controller is used as feedback to judge whether the parameters in the target operation instruction are the target parameters or not, so that the aim of enhancing the image data is fulfilled, the effect after enhancing the image data is better, the parameters of data enhancement operation corresponding to the whole image data set are prevented from being trained, and the efficiency of determining the target parameters is improved.
Fig. 3 shows a flow chart of a method for enhancing image data according to a further embodiment of the present invention, as shown in fig. 3, the method comprising:
308, enhancing the image data according to the target operation instruction to generate an image data sample;
and step 316, enhancing the image data according to the data enhancement operation and the target parameter.
In this embodiment, the accuracy of the image data sample is subjected to quantization processing by a reinforcement learning algorithm, and the accuracy of the image data sample subjected to quantization processing is sent to the neural network controller, that is, the neural network controller fits parameters in the target operation instruction according to the accuracy information subjected to quantization processing, and determines the parameters in the target operation instruction as target parameters. The efficiency of determining the target parameters is improved by the cooperation of the reinforcement learning algorithm and the neural network controller.
Fig. 4 shows a flow chart of a method for enhancing image data according to a further embodiment of the present invention, as shown in fig. 4, the method comprising:
In this embodiment, the neural network controller is used to fit the parameters in the target operation instruction, and when the fitting is not achieved, the neural network controller is used to perform resampling in the discrete search space, that is, the neural network controller is used to perform multiple samplings in the discrete search space. The neural network controller carries out random sampling for the first time in a discrete search space, generates a first image data sample according to a sampled first target operation instruction, detects the accuracy of the first image data sample output by the network, inputs the accuracy of the first image data sample into the neural network controller as a reward signal, carries out fitting for the first time on parameters in the first target operation instruction, and determines the parameters in the first target operation instruction as target parameters when fitting is achieved; when the fitting is not achieved, the neural network controller adjusts the sampling direction according to the first fitting result, second sampling is carried out in the discrete search space, a second image data sample is generated according to a second sampled target operation instruction, the accuracy of the second image data sample output by the sub-detection network is used as an incentive signal to be input to the neural network controller, the neural network controller carries out second fitting on parameters in the second target operation instruction, and when the fitting is achieved, the parameters in the second target operation instruction are determined to be target parameters; when the fitting is not achieved, the process is repeated. The neural network controller adjusts the sampling direction according to the fitting result, and then iterative training of parameters in the target operation instruction is achieved, so that the efficiency of determining the target operation parameters is higher, the determined target parameters are optimal solutions or approximate optimal solutions infinitely close to the optimal solutions, and the requirement of data enhancement is met.
In the above embodiment, further, the data enhancement operation includes: any one or more of a cropping operation, a geometric transformation operation, a rotation operation, an automatic contrast operation, an inversion operation, a color equalization operation, a color transformation operation, a luminance transformation operation, a sharpness transformation operation, and a histogram equalization operation.
In this embodiment, the data enhancement operation includes any one or more of ten operations, i.e., a cropping operation, a geometric transformation operation, a rotation operation, an automatic contrast operation, a reversal operation, a color equalization operation, a color transformation operation, a brightness transformation operation, a sharpness transformation operation, and a histogram equalization operation, and the image data set is enhanced according to different data enhancement operations, so that the generated image data set is larger in number, and better meets the requirements of a reinforcement learning algorithm, and reasonable target parameters are determined.
In the above embodiment, further, when the data enhancement operation includes X data enhancement operations, a seed enhancement strategies are determined, each seed enhancement strategy includes b data enhancement operations, where a × b ═ X, and a, b, and X are positive integers; the discrete space size of any data enhancement operation is: ((N X P M)b)aWhere N is the number of operations of the data enhancement operation, P is the discrete value of the probability of use of any kind of data enhancement operation, and M is the discrete value of the strength of any kind of data enhancement operation.
In this embodiment, the size of the discrete space is determined according to the data enhancement operation, for example, when the data enhancement operation includes all 10 data enhancement operations described above, 5 seed enhancement strategies are determined, each seed enhancement strategy including 2 data enhancement operations, and then the discrete space rotated by 10 degrees has a size ((10 × 11 × 10)2)5Where 10 is the number of operations of the data enhancement operation, 11 is the discrete value of the probability of use of the rotation operation, and 10 is the discrete value of the intensity of the rotation operation.
In the above embodiment, further, the reinforcement learning algorithm is a PPO algorithm, and a formula of the PPO algorithm is as follows:
wherein, θ and θ 'are sub-enhancement strategies, KL (θ, θ') is the KL divergence of the two sub-enhancement strategies, β is a penalty coefficient, and β is 1 e-4.
In the embodiment, the reinforcement learning algorithm is a PPO algorithm, penalty compensation is performed according to a formula, the corresponding accuracy of the obtained image data sample is higher, and the target parameter is more accurate.
An embodiment of the second aspect of the present invention provides an image data enhancement system, a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the image data enhancement method according to any one of the above items when executing the computer program. Therefore, the image data enhancement system comprises all the beneficial effects of the image data enhancement method of any one of the technical schemes.
According to a third aspect of the present invention, a computer-readable storage medium is proposed, on which a computer program is stored, which computer program, when being executed by a processor, performs the steps of the image data enhancement method of any of the above. Therefore, the computer readable storage medium includes all the advantages of the image data enhancement method of any one of the above technical solutions.
According to a fourth aspect of the present invention, there is provided a yard coverage detection method comprising the image data enhancement method of any one of the above, wherein the image data is acquired yard coverage image data. Therefore, the storage yard coverage detection method has all the beneficial effects of the image data enhancement method of any one of the technical schemes.
In order to implement the environmental protection law of the people's republic of China and the environmental protection law of the solid waste pollution of the people's republic of China to prevent and treat the secondary pollution of the storage and disposal sites of the general industrial solid waste, the China environmental protection department sets a pollution control standard of the storage and disposal sites of the general industrial solid waste (GB18599-2001), and the standard specifies that dust screens or dust cloths should be arranged on material piles, residue soil piles, waste residues, building materials and the like which are easy to generate dust emission, and spray treatment is carried out if necessary. Therefore, the detection work of whether various materials in the storage yard are covered is important.
The storage yard data collected by the actual camera has the following characteristics: 1) the material of various material piles is different from that of coal piles, muck piles, waste residues, building materials and the like, so that the material piles are different in size and shape; 2) the dust-proof cloth of the dust-proof net is not in a unified standard and has blue, black and green colors; 3) the acquired data is modulated by camera parameters, shooting angles and shooting environments of a camera, so that a large amount of manpower and time are needed to label covered areas and uncovered areas in a shot picture one by one.
The yard coverage detection method provided by the embodiment collects yard coverage images at preset time intervals, collects certain time, such as half a month, the yard coverage image data collected in this way contains image data under different light conditions and different weather conditions, and has better diversity, and meanwhile, the preset time interval can be 30-60 minutes, the coverage state cannot change too much in a short time, so that the working time of the camera can be reduced, the quantity of the collected image data is also avoided to be huge, the similarity is too high, and the parameter training of the image data enhancement operation is not facilitated.
In addition, according to the yard coverage detection method provided by this embodiment, after the yard coverage image data is acquired, the partial image data is manually labeled, and after the manual labeling, the image data is subjected to data image enhancement operation, so that label coordinate information corresponding to the image data is generated after the image data is enhanced.
In the yard coverage detection method, in order to select a suitable data enhancement operation as much as possible or to perform a certain limitation on the data enhancement operation, if it is not feasible to perform a color transformation operation on an image with white as a background and a single color, when a rotation operation is used, the rotation degree exceeding 90 ° does not have any physical significance to reality, and in addition, for example, the rotation operation, the effect of the rotation of 5 degrees and the rotation of 10 degrees on the data enhancement is similar, but the influence of the rotation of 10 degrees and the rotation of 20 degrees on the image data, on the reinforcement learning algorithm and the final detection result is great, and the yard image of some scenes can be completely understood as another image after the rotation of 20 degrees, so the application sets the ten data enhancement operations according to the actual situation.
It is worth mentioning that when the cameras for collecting image data are from different storage yards, the cameras used in different storage yards are configured differently, the shooting positions are different, and if a new scene is changed, parameters of data enhancement operation need to be retrained.
In the description herein, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance unless explicitly stated or limited otherwise; the terms "connected," "mounted," "secured," and the like are to be construed broadly and include, for example, fixed connections, removable connections, or integral connections; may be directly connected or indirectly connected through an intermediate. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
In the description herein, the description of the terms "one embodiment," "some embodiments," "specific embodiments," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes will occur to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. An image data enhancement method, comprising:
acquiring a data enhancement operation instruction, wherein the data enhancement operation instruction comprises a data enhancement operation and parameters of the data enhancement operation;
training the parameters of the data enhancement operation, and determining the target parameters of the data enhancement operation;
and enhancing the image data according to the data enhancement operation and the target parameters.
2. The image data enhancement method according to claim 1, wherein the training of the parameters of the data enhancement operation to determine the target parameters of the data enhancement operation specifically comprises:
determining a discrete search space according to the data enhancement operation instruction;
sampling in the discrete search space through a neural network controller to obtain a target operation instruction;
enhancing the image data according to the target operation instruction to generate an image data sample;
generating a sub-detection network according to the image data sample, and outputting the accuracy of the image data sample through the sub-detection network;
and fitting the parameters in the target operation instruction through a neural network controller according to the accuracy of the image data sample, and determining the parameters in the target operation instruction as target parameters when the parameters are fitted.
3. The method of claim 2, wherein the outputting the accuracy of the image data samples via the sub-detection network further comprises:
and preprocessing the accuracy of the image data sample through a reinforcement learning algorithm, and sending the accuracy of the preprocessed image data sample to the neural network controller.
4. The image data enhancement method according to claim 2, further comprising:
and fitting the parameters in the target operation instruction through the neural network controller, and sampling again in the discrete search space through the neural network controller when the fitting is not achieved.
5. The image data enhancement method of any one of claims 2 to 4, wherein the data enhancement operation comprises: any one or more of a cropping operation, a geometric transformation operation, a rotation operation, an automatic contrast operation, an inversion operation, a color equalization operation, a color transformation operation, a luminance transformation operation, a sharpness transformation operation, and a histogram equalization operation.
6. The image data enhancement method of claim 5, wherein when the data enhancement operation comprises X data enhancement operations,
determining a seed enhancement strategies, wherein each seed enhancement strategy comprises b data enhancement operations, a and b are X, and a, b and X are positive integers;
the discrete space size of any one of the data enhancement operations is: ((N X P M)b)a，
Wherein N is the number of operations of the data enhancement operation, P is the discrete value of the probability of use of any of the data enhancement operations, and M is the discrete value of the strength of any of the data enhancement operations.
7. The image data enhancement method of claim 3, wherein the reinforcement learning algorithm is a PPO algorithm, and the formula of the PPO algorithm is as follows:
wherein, θ and θ 'are sub-enhancement strategies, KL (θ, θ') is the KL divergence of the two sub-enhancement strategies, β is a penalty coefficient, and β is 1 e-4.
8. An image data enhancement system, comprising: a memory storing a computer program and a processor performing the steps of the data enhancement method of any one of claims 1 to 7 when the computer program is executed.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the data enhancement method according to any one of claims 1 to 7.
10. A yard coverage detection method comprising the image data enhancement method of any one of claims 1 to 7, wherein the image data is acquired yard coverage image data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010601516.7A CN111899186A (en) | 2020-06-29 | 2020-06-29 | Image data enhancement method, system, storage medium and storage yard coverage detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010601516.7A CN111899186A (en) | 2020-06-29 | 2020-06-29 | Image data enhancement method, system, storage medium and storage yard coverage detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111899186A true CN111899186A (en) | 2020-11-06 |
Family
ID=73206475
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010601516.7A Pending CN111899186A (en) | 2020-06-29 | 2020-06-29 | Image data enhancement method, system, storage medium and storage yard coverage detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111899186A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112613553A (en) * | 2020-12-18 | 2021-04-06 | 中电金信软件有限公司 | Picture sample set generation method and device, computer equipment and storage medium |
| CN113469078A (en) * | 2021-07-07 | 2021-10-01 | 西安电子科技大学 | Hyperspectral image classification method based on automatic design long-time and short-time memory network |
| CN114612859A (en) * | 2022-02-25 | 2022-06-10 | 交通运输部天津水运工程科学研究所 | Intelligent detection method for ore stacking tarpaulin of non-specialized wharf |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190046148A1 (en) * | 2017-08-11 | 2019-02-14 | Siemens Healthcare Gmbh | Method for analyzing image data from a patient after a minimally invasive intervention, analysis apparatus, computer program and electronically readable data storage medium |
| CN110033035A (en) * | 2019-04-04 | 2019-07-19 | 武汉精立电子技术有限公司 | A kind of AOI defect classification method and device based on intensified learning |
| CN110378419A (en) * | 2019-07-19 | 2019-10-25 | 广东浪潮大数据研究有限公司 | A kind of image set extending method, device, equipment and readable storage medium storing program for executing |
| CN110795623A (en) * | 2019-10-10 | 2020-02-14 | 深圳供电局有限公司 | Image enhancement training method and system and computer readable storage medium |
| CN110796248A (en) * | 2019-08-27 | 2020-02-14 | 腾讯科技（深圳）有限公司 | Data enhancement method, device, equipment and storage medium |
| CN110807109A (en) * | 2019-11-08 | 2020-02-18 | 北京金山云网络技术有限公司 | Data enhancement strategy generation method, data enhancement method and device |
-
2020
- 2020-06-29 CN CN202010601516.7A patent/CN111899186A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190046148A1 (en) * | 2017-08-11 | 2019-02-14 | Siemens Healthcare Gmbh | Method for analyzing image data from a patient after a minimally invasive intervention, analysis apparatus, computer program and electronically readable data storage medium |
| CN110033035A (en) * | 2019-04-04 | 2019-07-19 | 武汉精立电子技术有限公司 | A kind of AOI defect classification method and device based on intensified learning |
| CN110378419A (en) * | 2019-07-19 | 2019-10-25 | 广东浪潮大数据研究有限公司 | A kind of image set extending method, device, equipment and readable storage medium storing program for executing |
| CN110796248A (en) * | 2019-08-27 | 2020-02-14 | 腾讯科技（深圳）有限公司 | Data enhancement method, device, equipment and storage medium |
| CN110795623A (en) * | 2019-10-10 | 2020-02-14 | 深圳供电局有限公司 | Image enhancement training method and system and computer readable storage medium |
| CN110807109A (en) * | 2019-11-08 | 2020-02-18 | 北京金山云网络技术有限公司 | Data enhancement strategy generation method, data enhancement method and device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112613553A (en) * | 2020-12-18 | 2021-04-06 | 中电金信软件有限公司 | Picture sample set generation method and device, computer equipment and storage medium |
| CN112613553B (en) * | 2020-12-18 | 2022-03-08 | 中电金信软件有限公司 | Picture sample set generation method and device, computer equipment and storage medium |
| CN113469078A (en) * | 2021-07-07 | 2021-10-01 | 西安电子科技大学 | Hyperspectral image classification method based on automatic design long-time and short-time memory network |
| CN114612859A (en) * | 2022-02-25 | 2022-06-10 | 交通运输部天津水运工程科学研究所 | Intelligent detection method for ore stacking tarpaulin of non-specialized wharf |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3631754B1 (en) | Image processing apparatus and method | |
| US7653238B2 (en) | Image filtering based on comparison of pixel groups | |
| CN111899186A (en) | Image data enhancement method, system, storage medium and storage yard coverage detection method | |
| WO2022000397A1 (en) | Low-illumination image enhancement method and apparatus, and computer device | |
| Fan et al. | No reference image quality assessment based on multi-expert convolutional neural networks | |
| US10417772B2 (en) | Process to isolate object of interest in image | |
| CN107705254B (en) | City environment assessment method based on street view | |
| CN109242787A (en) | It paints in a kind of assessment of middle and primary schools' art input method | |
| CN106960196B (en) | Industrial video small number recognition method based on template matching and SVM | |
| CN109740553A (en) | A kind of image, semantic segmentation data screening method and system based on identification | |
| Zhuang et al. | Image enhancement by deep learning network based on derived image and retinex | |
| Hori et al. | A real-time multi face detection technique using positive-negative lines-of-face template | |
| Braik et al. | Enhancement of satellite images based on CLAHE and augmented elk herd optimizer | |
| CN116385333A (en) | Laser cleaning robot cleaning effect evaluation method and system | |
| CN117237779B (en) | Image recognition method and system for combined analysis of visible light images and infrared images | |
| DE102012007838B4 (en) | High dynamic range image acquisition with multiple exposure | |
| Flotzinger et al. | dacl1k: Real-world bridge damage dataset putting open-source data to the test | |
| CN117037075A (en) | Engineering detection method and system based on image processing | |
| CN109285166B (en) | Overlapping and conglutinating chromosome automatic segmentation method based on full convolution network | |
| CN116091436A (en) | Image sample collection method and device applied to agricultural field | |
| CN113472998B (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| DE112022002910T5 (en) | REINFORCEMENT LEARNING BASED SYSTEM FOR TUNING CAMERA PARAMETERS TO IMPROVE ANALYTICS | |
| CN112581485A (en) | Imaging-free edge information detection method and system | |
| Wang et al. | Optimized contrast enhancements to improve robustness of visual tracking in a SLAM relocalisation context | |
| DE112021008046T5 (en) | Computer-aided photography in low light conditions |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20201106 |
|
| WD01 | Invention patent application deemed withdrawn after publication |