JP2018172051A - Parking support device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce a processing amount while suppressing deterioration of identification accuracy of a parkable area in a parking support device for supporting parking of a vehicle. An extraction unit 62 extracts a license plate from an image around a vehicle captured by an image pickup device 10 mounted on a vehicle 100. Based on the license plate extracted by the extraction unit 62, the exclusion unit 66 excludes from the image around the vehicle a region in which another vehicle to which the license plate is attached is arranged. The detection unit 68 detects the feature points in the image excluding the exclusion region in the exclusion unit 66. The parkable area specifying unit 70 identifies an area in which the vehicle 100 can be parked from an image excluding the excluded area in the exclusion unit 66 based on the feature points detected by the detection unit 68. [Selection diagram] Fig. 1

Description

  The present invention relates to a parking assistance device that identifies an area where parking is possible.

  In order to support safe driving of a vehicle, it is required to grasp the positional relationship with other vehicles around the host vehicle from the image. Therefore, a license plate is extracted from the image, and the positional relationship with other vehicles is calculated from the character size (see, for example, Patent Document 1).

JP 2005-30925 A

  When assisting parking of a vehicle as a safe driving of the vehicle, an area where parking is possible is specified from the image. Here, when an area where obstacles such as other vehicles are not arranged is specified for the entire image, the processing amount is large and the processing time is long. On the other hand, if the amount of processing is reduced in order to shorten the processing time, the accuracy of specifying a region where no obstacle is arranged, that is, a region where parking is possible, deteriorates.

  The present invention has been made in view of such a situation, and an object of the present invention is to provide a technology for reducing the amount of processing while suppressing deterioration in the accuracy of specifying a parking area in a parking assistance device that supports parking of a vehicle. It is to provide.

  In order to solve the above problems, a parking assist device according to an aspect of the present invention includes an extraction unit that extracts a license plate from an image around a vehicle captured by an imaging device mounted on the vehicle, and a number extracted by the extraction unit. Based on the plate, the area where other vehicles with the license plate are placed is the excluded area, and the excluded part that is excluded from the image around the vehicle and the feature points in the image that excludes the excluded area in the excluded part are detected And a parking area specifying unit that specifies an area in which the vehicle can be parked from an image in which the exclusion area is excluded by the exclusion unit based on the feature points detected by the detection unit.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, in the parking assistance apparatus which assists parking of a vehicle, a processing amount can be reduced, suppressing the deterioration of the specific precision of the area | region which can be parked.

1 is a diagram illustrating a configuration of a vehicle according to a first embodiment. It is a figure which shows the image used as the object of a process in the input part of FIG. 1, and an extraction part. It is a figure which shows the detection process of the feature point used as the comparison object of the parking assistance apparatus of FIG. FIGS. 4A and 4B are diagrams showing processing areas set in the setting unit shown in FIG. It is a figure which shows the data structure of the table memorize | stored in the setting part of FIG. It is a figure which shows the image used as the object of a process in the other vehicle specific part of FIG. It is a figure which shows the image used as the object of a process in the exclusion part of FIG. 1, and a detection part. It is a flowchart which shows the specific procedure by the parking assistance apparatus of FIG. It is a figure which shows the exclusion area | region set in the setting part which concerns on Example 2. FIG. It is a figure which shows the image used as the object of a process in the exclusion part which concerns on Example 2, and a detection part. 12 is a flowchart illustrating a specific procedure by the parking assist device according to the second embodiment. It is a figure which shows the image used as the object of a process in the exclusion part and detection part which concern on a modification.

Example 1
Before describing the present invention in detail, an outline will be described. An embodiment of the present invention relates to a parking support apparatus that supports parking of a vehicle. In particular, the parking assist device executes a process for specifying a parking area in a parking lot or the like. In this process, (1) a feature point in an image around the vehicle imaged by an imaging device mounted on the vehicle is detected, and (2) the position of an obstacle is specified by estimating the three-dimensional position of the feature point. (3) A portion where no obstacle exists is specified as a parking area. However, the number of feature points in the entire image is generally large, and the processing amount for specifying the position of the obstacle increases, so that the processing time becomes long. Further, when the amount of processing increases, there is a possibility that processing cannot be performed in an apparatus mounted on a vehicle.

  In order to cope with these, so far, the processing amount has been reduced by thinning out the feature points. However, by thinning out the feature points, the estimation accuracy of the three-dimensional position is deteriorated, and the identification accuracy of the parking area is also deteriorated. For example, an error of about 50 cm may occur with respect to a three-dimensional position. Therefore, it is required to reduce the processing amount while suppressing the deterioration of the specific accuracy. The parking assistance device in the present embodiment extracts the number plate included in the image by pattern matching, and specifies the position of the other vehicle (the region where the other vehicle exists) based on the position of the number plate. Following this, the parking assistance device identifies the part where no obstacle exists as a parking area by extracting feature points in the remaining part excluding other vehicles from the image and estimating the three-dimensional position. To do.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, each Example described below is an example, and this invention is not limited by these Examples.

  FIG. 1 shows a configuration of a vehicle 100 according to the first embodiment. The vehicle 100 includes an imaging device 10, a position information acquisition unit 12, a parking assist device 20, a brake sensor 30, a steering angle sensor 32, a vehicle speed sensor 34, a steering actuator 36, a brake 38, and an accelerator 40. The parking assist device 20 includes a parking area search unit 50, a target position determination unit 52, a route generation unit 54, and a vehicle control unit 56. The parking area search unit 50 includes an input unit 60, an extraction unit 62, an other vehicle estimation unit 64, an exclusion unit 66, a detection unit 68, and a parking area identification unit 70, and the other vehicle estimation unit 64 includes a setting unit 80 and other vehicles. A specifying unit 82 is included.

  The imaging device 10 is a monocular camera mounted in front of or behind the vehicle 100. A plurality of imaging devices 10 may be mounted on the front, rear, left and right of the vehicle 100. The imaging device 10 repeatedly captures an image around the vehicle 100 and outputs image data of the captured image (hereinafter also referred to as “image”) to the parking assist device 20. The position information acquisition unit 12 acquires the current position of the vehicle 100 from a GNSS (Global Navigation Satellite System (s)) receiver. Further, the position information acquisition unit 12 may acquire the current position using a steering angle, a wheel speed pulse, or the like. The position information acquisition unit 12 outputs the current position to the parking assistance device 20.

  The input unit 60 receives an image from the imaging device 10. FIG. 2 shows an image 200 to be processed in the input unit 60 and the extraction unit 62. The image 200 includes a first parking area 210a to a fifth parking area 210e, collectively referred to as a parking area 210, a first other vehicle 212a to a fourth other vehicle 212d, collectively referred to as an other vehicle 212, a wall 214, and a tree 216. It is. Here, the first other vehicle 212a is parked in the first parking area 210a, the second other vehicle 212b is parked in the second parking area 210b, and the third other vehicle 212c is parked in the fourth parking area 210d. The fourth other vehicle 212d is parked in the fifth parking area 210e. On the other hand, since the other vehicle 212 is not parked in the third parking area 210c, the vehicle 100 can be parked. Therefore, the goal of the parking area search unit 50 is to specify the third parking area 210c as a parking area by performing image processing on the image 200.

  As described above, as described above, (1) processing for detecting feature points in the image 200, (2) processing for specifying the position of an obstacle by estimating the three-dimensional position of the feature points, and (3) obstacle The process of specifying the part where no is present is executed in order. FIG. 3 shows a process for detecting the feature point 220 to be compared with the parking assistance device 20. Here, a plurality of feature points 220 detected from the image 200 by the process (1) are shown. Since the entire image 200 is a processing target, the number of feature points 220 is increased. Therefore, the processing amount of (2) becomes large, and the processing for the entire image 200 is not realistic. In order to cope with this, the parking area search unit 50 executes the following process. Returning to FIG.

  The extraction unit 62 extracts a license plate from the image 200 received by the input unit 60. Pattern matching is used to extract the license plate. In pattern matching, a position similar to the template is searched from the image 200 using the license plate as a template. Such pattern matching is also called template matching. Furthermore, the extraction part 62 also acquires the color of the extracted license plate and the characters indicated on the license plate. In particular, the characters shown on the license plate correspond to classification numbers. Since a known technique may be used for this, description thereof is omitted here. When extracting the license plate, the extraction unit 62 outputs information on the position (including dimensions), color, and classification number of the extracted license plate to the setting unit 80.

  The setting unit 80 sets a processing region 232 surrounding the license plate at the position received from the extraction unit 62. In order to describe the processing region 232, FIGS. 4A to 4B and FIG. 5 are used here. FIGS. 4A and 4B show images in the processing area 232 set by the setting unit 80, and these correspond to areas where the other vehicle 212 included in the image 200 exists. In the extraction unit 62, the license plate 230 provided in the other vehicle 212 is extracted from the image 200, so the position of the license plate 230 is specified. In FIG. 4A, the width of the license plate 230 is indicated as “w”, and the height of the license plate 230 is indicated as “h”.

  The size of the width “w” and the height “h” in the license plate 230 included in the image 200 changes according to the distance from the vehicle 100 to the other vehicle 212. The setting unit 80 sets a length obtained by multiplying the width “w” in the license plate 230 included in the image 200 by “α21” as the width of the processing region 232, and sets the height “h” in the license plate 230 to “β21”. The length multiplied by “is set in the processing area 232. Furthermore, the setting unit 80 matches the center of the license plate 230 in the width direction with the center of the processing region 232 in the width direction, and the lower end of the processing region 232 is on the road surface (for example, the road surface where the tire of the vehicle 100 contacts). The processing area 232 is arranged on the image 200 so as to match. FIG. 4B will be described later, and the coefficients “α21” and “β21” of the processing region 232 will be described next.

  FIG. 5 shows a data structure of a table stored in the setting unit 80. Here, it is assumed that the other vehicle 212 is a mini vehicle, a small vehicle, or a normal vehicle. These differences are shown in “Vehicle Size”. The actual size of the license plate 230 is indicated by “number plate”. In the type of the other vehicle 212 described above, the actual width “w” and the actual height “h” of the license plate 230 are common. On the other hand, the coefficients “α21” and “β21” of the processing area 232 are determined for each type of the other vehicle 212. Further, the actual size “w · α21” of the entire width of the processing region 232 and the actual size “h · β21” of the height are determined for each type of the other vehicle 212. The coefficients “α21” and “β21” of the processing region 232 determined for each type of the other vehicle 212 are set so that the processing region 232 includes the shape of the outer edge of the other vehicle 212 of each type. Returning to FIG.

  When the color information received from the extraction unit 62 indicates yellow, the setting unit 80 specifies that the type of the vehicle 100 is “light vehicle”. In addition, when the color information received from the extraction unit 62 indicates white, the setting unit 80 specifies “small car” or “ordinary car” according to the classification number received from the extraction unit 62. The setting unit 80 acquires coefficients “α21” and “β21” corresponding to the specified type of the vehicle 100 from the table, and sets the processing region 232 using these. That is, the setting unit 80 changes the size of the processing region 232 according to the license plate 230 extracted by the extraction unit 62. When the extraction unit 62 extracts a plurality of license plates 230 from the image 200, the setting unit 80 sets a plurality of processing regions 232 so as to correspond to the respective license plates 230.

  The other vehicle specifying unit 82 detects feature points in the image 200 limited to the processing region 232 set in the setting unit 80. FIG. 6 shows an image 200 to be processed in the other vehicle specifying unit 82. Since four number plates 230 corresponding to the fourth other vehicle 212d are extracted from the first other vehicle 212a in the extraction unit 62, the first processing region 232a to the fourth processing region are set in the setting unit 80 so as to correspond to each. Four processing areas 232 up to 232d are set. The other vehicle specifying unit 82 detects feature points limited to each processing region 232. Since a known technique may be used for the process for detecting the feature points, the description is omitted here. On the other hand, the other vehicle specifying unit 82 does not execute processing for detecting feature points in a portion other than the processing region 232 in the image 200. As a result, the feature point 240 is detected only within the four processing regions 232. Thereby, the number of detected feature points 240 is reduced.

  The other vehicle specifying unit 82 estimates the three-dimensional position of each feature point 240 detected in the four processing regions 232. For the estimation of the three-dimensional position, for example, a known SfM (Structure from Motion) technique is used. The other vehicle specifying unit 82 specifies the other vehicle 212 by estimating the three-dimensional position. In FIG. 6, the position and shape for each of the first other vehicle 212a to the fourth other vehicle 212d are specified. The other vehicle identification unit 82 outputs information regarding the identified other vehicle 212 to the exclusion unit 66.

  Note that the processing region 232 set in the setting unit 80 may include a central portion 234 as shown in FIG. The central portion 234 is disposed at the center in the width direction of the processing region 232. At that time, the other vehicle specifying unit 82 omits the detection process of the feature point 240 in the central portion 234 of the processing region 232 set in the setting unit 80. As a result, the number of detected feature points 240 is further reduced.

  The exclusion unit 66 receives information related to the other vehicle 212 from the other vehicle specifying unit 82. The exclusion unit 66 sets an exclusion region 250 having the shape of the outer edge of the other vehicle 212 included in the received information in the image 200. FIG. 7 shows an image 200 to be processed in the exclusion unit 66 and the detection unit 68. The first exclusion region 250a is set to coincide with the outer edge of the first other vehicle 212a, and the second exclusion region 250b is set to coincide with the outer edge of the second other vehicle 212b. Further, the third exclusion region 250c is set so as to coincide with the outer edge of the third other vehicle 212c, and the fourth exclusion region 250d is set so as to coincide with the outer edge of the fourth other vehicle 212d. Note that the exclusion region 250 may have a size larger than the outer edge of the other vehicle 212. Thus, the exclusion area 250 is an area where the other vehicle 212 to which the license plate 230 is attached is present, and can be said to be an area where the other vehicle 212 specified by the other vehicle specifying unit 82 exists. Returning to FIG. The exclusion unit 66 outputs the image 200 excluding the exclusion region 250 to the detection unit 68.

  The detection unit 68 detects feature points in the image 200 in which the exclusion unit 66 excludes the exclusion region 250. The feature points are detected in the same manner as the other vehicle specifying unit 82. In FIG. 7, the feature point 260 is detected in the image 200 other than the excluded area 250. Since the feature points 260 are not detected in the exclusion region 250, the number of detected feature points 260 is reduced.

  The parking area specifying unit 70 estimates the three-dimensional position of the feature point 260 detected by the detecting unit 68, similarly to the other vehicle specifying unit 82. The parking area specifying unit 70 specifies obstacles such as the wall 214 and the tree 216 by estimating the three-dimensional position. The parking area specifying unit 70 is an area other than the part where the obstacle is arranged in the image 200 from which the exclusion area 250 is excluded in the exclusion section 66 and has an area that allows the vehicle 100 to be parked. The parking area 270 is specified. More preferably, an area that is an area between adjacent exclusion areas 250 and has an area that allows the vehicle 100 to be parked is specified as the parking area 270.

  The target position determination unit 52 determines a target position for the parking area 270 specified by the parking area specification unit 70. The route generation unit 54 generates a route from the current position of the vehicle 100 acquired by the position information acquisition unit 12 to the target position determined by the target position determination unit 52. In addition, when the vehicle 100 needs to be turned back from the current position to the target position, the route generation unit 54 generates a route including the turn back.

  The brake sensor 30 detects the position of the brake pedal and outputs the detected position to the parking assist device 20. The steering angle sensor 32 detects the steering angle of the wheel and outputs the detected steering angle to the parking assist device 20. The vehicle speed sensor 34 detects the speed of the vehicle 100 and outputs the detected vehicle speed to the parking assist device 20. The steering actuator 36 drives the steering according to the control of the parking assist device 20. The brake 38 decelerates the vehicle 100 according to the control of the parking assist device 20. The accelerator 40 accelerates the vehicle 100 according to the control of the parking assist device 20.

  The vehicle control unit 56 receives the current position from the position information acquisition unit 12, the position of the brake pedal from the brake sensor 30, the steering angle from the steering angle sensor 32, and the speed from the vehicle speed sensor 34. Based on the received information, the vehicle control unit 56 controls the steering actuator 36, the brake 38, and the accelerator 40 so that the vehicle 100 reaches the target position along the route generated by the route generation unit 54. . This corresponds to controlling the automatic parking of the vehicle 100 so that the vehicle 100 is parked in the parking area 270 along the route.

  This configuration can be realized in terms of hardware by a CPU, memory, or other LSI of any computer, and in terms of software, it can be realized by a program loaded in the memory, but here it is realized by their cooperation. Draw functional blocks. However, it will be understood by those skilled in the art that these functional blocks can be realized in various forms by hardware alone or by a combination of hardware and software.

  Operation | movement of the parking assistance apparatus 20 by the above structure is demonstrated. FIG. 8 is a flowchart showing a specific procedure by the parking assistance device 20. The imaging device 10 captures an image 200 around the vehicle 100 (S10). The extraction unit 62 extracts the license plate 230 included in the image 200 (S12). The setting unit 80 sets a predetermined processing area 232 surrounding the license plate 230 (S14). The other vehicle specifying unit 82 detects the feature point 240 in the processing area 232 (S16), and specifies the other vehicle 212 (S18). The exclusion part 66 sets the exclusion area | region 250 corresponding to the outer edge of the other vehicle 212 (S20). The detecting unit 68 detects the feature point 260 in the image 200 excluding the excluded region 250 (S22). If there is an empty area in which the vehicle 100 can be parked between adjacent exclusion areas 250 (Y in S24), the parking area specifying unit 70 specifies the empty area as the parking area 270 (S26). If there is no empty area in which vehicle 100 can be parked between adjacent exclusion areas 250 (N in S24), the process ends.

  According to the first embodiment, the feature points in the image excluding the excluded region are detected, so that the number of feature points can be reduced. Moreover, since the number of feature points is reduced, the amount of processing can be reduced. Further, since the processing amount is reduced, the processing time can be shortened. Moreover, since the area | region which this vehicle can park is identified based on the feature point in the image which excluded the exclusion area | region, the deterioration of specific accuracy can be suppressed. In addition, since feature points in an image around the vehicle are detected only in a predetermined processing area and other vehicles are identified based on the detected feature points, the processing amount is suppressed while suppressing deterioration in the identification accuracy of the other vehicles. Can be reduced. Further, since the feature point detection process is omitted in the central portion of the processing area, the processing amount can be reduced. Moreover, since the size of the processing area is changed according to the number plate, the identification accuracy of the other vehicle can be improved.

(Example 2)
Next, Example 2 will be described. As in the first embodiment, the second embodiment relates to a parking assistance device that executes a process for specifying a region where a vehicle can be parked in a parking lot or the like. The second embodiment aims to further simplify the processing of the first embodiment. The parking assist device according to the first embodiment specifies the position of the other vehicle (the shape of the outer edge) in the processing region set based on the position of the license plate, and sets the excluded region at the position of the specified other vehicle (the shape of the outer edge). It is set. On the other hand, the parking assistance apparatus according to the second embodiment sets the exclusion region without specifying the position of the other vehicle (the shape of the outer edge) based on the position of the license plate. The parking assistance apparatus in Example 2 is the same type as FIG. Here, it demonstrates centering on the difference with Example 1. FIG.

  The setting unit 80 sets a region having a predetermined size as an exclusion region 250 that is arranged so as to surround the license plate 230 extracted by the extraction unit 62. In order to explain the exclusion region 250, FIGS. 9 and 5 are used here. FIG. 9 shows an exclusion area 250 set in the setting unit 80 according to the second embodiment, which corresponds to the other vehicle 212 included in the image 200. The setting unit 80 sets a length obtained by multiplying the width “w” in the license plate 230 included in the image 200 by “α22” as the width of the exclusion region 250, and sets the height “h” in the license plate 230 to “β22”. The length multiplied by “is set in the exclusion area 250. Furthermore, the setting unit 80 matches the center in the width direction of the license plate 230 with the center in the width direction of the exclusion region 250 and sets the lower end of the exclusion region 250 to the road surface (for example, the road surface where the tire of the vehicle 100 contacts). The exclusion region 250 is arranged on the image 200 so as to match. As shown in FIG. 5, the coefficients “α22” and “β22” of the exclusion region 250 are determined for each type of the other vehicle 212. Further, the actual size “w · α22” of the entire width of the exclusion region 250 and the actual size “h · β22” of the height are determined for each type of the other vehicle 212. The coefficients “α22” and “β22” of the exclusion region 250 determined for each type of the other vehicle 212 are set so that the size of the exclusion region 250 is approximately the same as or smaller than the shape of the outer edge of the other vehicle 212 of each type. Is done. Returning to FIG.

  As in the first embodiment, the setting unit 80 uses the color information received from the extraction unit 62 and the classification number received from the extraction unit 62 as the type of the vehicle 100 as “light vehicle” or “small vehicle” Specify "ordinary vehicle". The setting unit 80 acquires coefficients “α22” and “β22” corresponding to the specified type of the vehicle 100 from the table, and sets the exclusion region 250 using these. That is, the setting unit 80 changes the size of the exclusion region 250 according to the license plate 230 extracted by the extraction unit 62. When the extraction unit 62 extracts a plurality of license plates 230 from the image 200, the setting unit 80 sets a plurality of exclusion areas 250 so as to correspond to the respective license plates 230.

  In the parking assistance device 20 according to the second embodiment, the other vehicle specifying unit 82 is omitted from the configuration of the parking assistance device 20 according to the first embodiment. The exclusion unit 66 receives the image 200 in which the exclusion region 250 is set by the setting unit 80. FIG. 10 illustrates an image 200 to be processed in the exclusion unit 66 and the detection unit 68 according to the second embodiment. The first excluded area 250a is set to match the first other vehicle 212a, and the second excluded area 250b is set to match the second other vehicle 212b. Further, the third excluded area 250c is set so as to coincide with the third other vehicle 212c, and the fourth excluded area 250d is set so as to coincide with the fourth other vehicle 212d. Thus, it can be said that the exclusion area 250 is an area arranged so as to surround the license plate 230 and an area having a predetermined size. As described above, the size of the exclusion region 250 is changed according to the license plate 230 extracted by the extraction unit 62. Returning to FIG. The exclusion unit 66 outputs the image 200 excluding the exclusion region 250 to the detection unit 68. Subsequent processing is the same as that in the first embodiment, and thus description thereof is omitted here.

  Operation | movement of the parking assistance apparatus 20 by the above structure is demonstrated. FIG. 11 is a flowchart illustrating a specific procedure performed by the parking assistance apparatus 20 according to the second embodiment. The imaging device 10 captures an image 200 around the vehicle 100 (S50). The extraction unit 62 extracts the license plate 230 included in the image 200 (S52). The setting unit 80 sets a predetermined exclusion area 250 surrounding the license plate 230 (S54). The detecting unit 68 detects the feature point 260 in the image 200 excluding the excluded region 250 (S56). If there is an empty area where the vehicle 100 can be parked between the adjacent exclusion areas 250 (Y in S58), the parking area specifying unit 70 specifies the empty area as the parking area 270 (S60). If there is no vacant area in which vehicle 100 can be parked between adjacent exclusion areas 250 (N in S58), the process ends.

  According to the second embodiment, since an area of a predetermined size is included as an exclusion area so as to surround the extracted license plate, the processing amount can be reduced. In addition, since an area of a predetermined size is set as an exclusion area so as to surround the extracted license plate, the processing can be simplified. Further, since the size of the exclusion area is changed according to the license plate, the processing amount can be reduced.

  In the first and second embodiments, the exclusion area is identified using the license plate information extracted from the captured image, and the vehicle can be parked based on the feature points in the image excluding the identified exclusion area. An area is specified. However, the present invention is not limited to this. For example, as shown in FIG. 12, the limited area 280 is defined so as to include all of the excluded areas 250 (first excluded area 250a to fourth excluded area 250d) specified using the license plate information. Alternatively, an area where the vehicle can be parked may be specified based on the feature points in the image in which the excluded area 250 is excluded from the defined limited area 280. In this case, the processing amount can be further reduced. For example, the limited region 280 may be defined to include a region where the other vehicle 212 is estimated to exist by estimating the three-dimensional position in a state where the feature points are thinned out.

  In the second embodiment, the size of the exclusion area 250 is changed according to the extracted license plate 230. However, the present invention is not limited to this. For example, the exclusion region 250 may be set to a predetermined fixed size regardless of the type of other vehicle. Thereby, the process at the time of specifying the exclusion area | region 250 can be simplified.

  In the second embodiment, the exclusion area 250 is set for each other vehicle 212 and is excluded from the image 200. However, the present invention is not limited to this. For example, the distance (interval) between the third exclusion region 250c and the fourth exclusion region 250d located adjacent thereto is estimated from the position of the license plate 230, and the distance (interval) is predetermined. If it is less than this value (for example, less than the width of the vehicle 100), the region between the two exclusion regions may be set as the exclusion region. Thereby, the processing amount can be further reduced.

  One embodiment of the present invention is as follows. A parking assist device according to an aspect of the present invention includes an extraction unit that extracts a license plate from an image around a vehicle imaged by an imaging device mounted on the vehicle, and the number plate based on the license plate extracted by the extraction unit. In an exclusion area that excludes an area around the vehicle where the other vehicle to which the plate is attached is placed, and a detection section that detects a feature point in the image that excludes the exclusion area in the exclusion section, and a detection section A parking area specifying unit that specifies an area in which the present vehicle can be parked from an image in which the exclusion area is excluded by the exclusion unit based on the detected feature point.

  According to this aspect, since the area where the present vehicle can be parked is specified based on the feature points in the image excluding the excluded area, the processing amount can be reduced while suppressing the deterioration of the specifying accuracy.

  A setting unit for setting a processing region surrounding the license plate extracted by the extraction unit, and a feature point in the image around the vehicle limited to the processing region set in the setting unit, and other based on the detected feature point You may further provide the other vehicle specific part which specifies this vehicle. The exclusion unit may set a region where another vehicle identified by the other vehicle identification unit is arranged as an exclusion region. In this case, the feature points in the image around the vehicle are detected only in the processing region, and other vehicles are identified based on the detected feature points, so that the processing can be performed while suppressing the deterioration of the identification accuracy of the other vehicles. The amount can be reduced.

  The other vehicle specifying unit may omit the feature point detection process in the central portion of the processing region set in the setting unit. In this case, since the feature point detection process is omitted in the central portion of the processing area, the processing amount can be reduced.

  The setting unit may change the size of the processing region according to the license plate extracted by the extraction unit. In this case, since the size of the processing region is changed according to the number plate, it is possible to improve the identification accuracy of other vehicles.

  The exclusion unit may be a region that is arranged so as to surround the license plate extracted by the extraction unit and that has a predetermined size. In this case, since an area of a predetermined size is set as an exclusion area so as to surround the extracted license plate, the processing amount can be reduced.

  The exclusion unit may change the size of the exclusion region according to the license plate extracted by the extraction unit. In this case, since the size of the exclusion area is changed according to the license plate, the processing amount can be reduced.

  In the above, this invention was demonstrated based on the Example. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to each of those constituent elements or combinations of processing processes, and such modifications are also within the scope of the present invention. .

  DESCRIPTION OF SYMBOLS 10 Imaging device, 12 Position information acquisition part, 20 Parking assistance apparatus, 30 Brake sensor, 32 Steering angle sensor, 34 Vehicle speed sensor, 36 Steering actuator, 38 Brake, 40 Axel, 50 Parking area search part, 52 Target position determination part, 54 route generation unit 56 vehicle control unit 60 input unit 62 extraction unit 64 other vehicle estimation unit 66 exclusion unit 68 detection unit 70 parking area identification unit 80 setting unit 82 other vehicle identification unit 100 vehicle.

Claims (6)

An extraction unit for extracting a license plate from an image around the vehicle imaged by an imaging device mounted on the vehicle;
Based on the license plate extracted in the extraction unit, as an exclusion region the region where other vehicles attached with the license plate are arranged, an exclusion unit that excludes from the image around the vehicle,
A detection unit for detecting a feature point in the image excluding the exclusion region in the exclusion unit;
Based on the feature points detected in the detection unit, a parking area specifying unit that specifies an area where the vehicle can be parked from an image excluding the exclusion area in the exclusion unit;
A parking assistance device comprising: A setting unit for setting a processing region surrounding the license plate extracted in the extraction unit;
In addition to detecting the feature points in the image around the vehicle limited to the processing region set in the setting unit, the vehicle further comprises an other vehicle specifying unit for specifying another vehicle based on the detected feature points,
The parking support device according to claim 1, wherein the exclusion unit sets an area where another vehicle specified by the other vehicle specification unit is arranged as an exclusion area.   The parking assistance device according to claim 2, wherein the other vehicle specifying unit omits a feature point detection process in a central portion of the processing region set in the setting unit.   The parking support device according to claim 2 or 3, wherein the setting unit changes the size of the processing area according to the license plate extracted by the extraction unit.   The parking support device according to claim 1, wherein the exclusion unit is a region arranged so as to surround the license plate extracted by the extraction unit, and a region having a predetermined size is defined as the exclusion region. .   The parking support device according to claim 5, wherein the exclusion unit changes a size of the exclusion region according to the license plate extracted by the extraction unit.

Images