TW201433774A - Vehicle specifications measurement processing device, vehicle specifications measuring method, and recording medium - Google Patents

Abstract

An image processing apparatus detects dimension-measurement feature points ♯1, ♯2, which are positioned on side faces of a vehicle, and a dimension measurement feature point ♯3, which is positioned on an under face of the vehicle and calculating the width of the vehicle by performing motion stereo processing on the dimension measurement feature points ♯1, ♯2 and ♯3. The dimension measurement feature points ♯1, ♯2 and ♯3, which are detected by the image processing apparatus, are detected so that the positions thereof have degrees of freedom on the side faces or the under face of the vehicle.

Description

Vehicle specification measurement processing device, vehicle specification measurement method, and recording medium

The present invention relates to a vehicle specification measurement processing device, a vehicle specification measurement method, and a recording medium, and more particularly to a vehicle specification measurement processing device, a vehicle specification measurement method, and a recording medium, and more particularly to a specification for measuring a vehicle (width, length) , height) devices and methods.

In a toll station such as a highway or a toll road, a vehicle type discriminating device for discriminating a vehicle type of the passing vehicle may be provided. Various methods have been used for vehicle type discrimination, but in recent years, an image of a vehicle has been studied, and the image to be captured is used for vehicle type discrimination. Since various information about the vehicle can be obtained from the image of the vehicle, the vehicle type discrimination using the vehicle image is useful for improving the accuracy of the vehicle type discrimination.

In the vehicle type discrimination using the image of the vehicle, the feature points are sometimes extracted from the image of the vehicle, and the vehicle type is determined based on the extracted feature points. For example, Non-Patent Document 1 (J Prokai, G Medioni, "3-D model based vehicle recognition," Applications of Computer Vision (WACV), 2009) discloses a technique in which registration is prepared in advance by using CG (computer Graphic, computer graphics) A database of templates created for each vehicle's vehicle learning feature points, using the matching of feature points from the image extraction and the template of the database to estimate the vehicle type. However, in this method, it is necessary to prepare a template for each vehicle type.

As another method of determining the vehicle type of the image of the vehicle, there is known a technique of calculating the specification (for example, width, length, and height) of the vehicle by image processing. The calculated vehicle specifications are used as information for vehicle type discrimination. For example, Patent Document 1 (Japanese Laid-Open Patent Publication No. Hei 11-86185) discloses a technique of generating a geometric correction image of a front side and a side surface of a vehicle by correcting an image of the captured vehicle, and The size of the vehicle is estimated based on the geometric correction image of the front and side. More specifically, the technique disclosed in Patent Document 1 detects the grounding position of the tire and the road surface, and calculates the position and the entering angle in the three-dimensional space of the vehicle based on the grounding position and the setting conditions of the imaging device. Correcting an image obtained by photographing the vehicle based on the calculated position and the angle of entry in the three-dimensional space of the vehicle, thereby generating a geometrically corrected image of the front and side of the vehicle, and correcting the image based on the generated geometric correction image The size of the vehicle.

Further, Non-Patent Document 2 (A. Lai, G. Fung and N. Yung, "Vehicle type classification from visual-based dimension estimation" IEEE, Intelligent Transportation Systems, pages 201-206, 2001) discloses that there are The technique of estimating the size of the vehicle by the size of the appearance of the vehicle. Specifically, the technique disclosed in Non-Patent Document 2 extracts an area corresponding to the shape of the vehicle in the image as a hexagonal area. Calculating the coordinates of the three vertices at the position near the road surface among the vertices of the hexagon, calculating the width and length of the vehicle based on the coordinates, and further calculating the vehicle according to the length of the hexagonal side corresponding to the height direction in the real space. height.

In the method of calculating the specification of the vehicle by image processing, it is preferable to calculate the vehicle by image processing that does not require the reliability of feature point extraction, that is, image processing that is robust to the detection of feature points. specification. When the specific feature points cannot be calculated, it is not preferable to calculate the specifications of the vehicle or the method of reducing the accuracy. For example, in the technique disclosed in Non-Patent Document 2, if it is impossible to detect three vertices corresponding to the vicinity of the road surface among the vertices of the hexagonal shape corresponding to the shape of the vehicle as the feature points, the specification of the vehicle cannot be calculated. A strong constraint on the detected feature points will result in an inaccurate measurement of the image processing using the specifications of the vehicle.

According to this background, it is required to provide a kind of robustness by detecting a feature point. A technique for calculating the specifications (for example, width, length, and height) of a vehicle by image processing.

[Previous Technical Literature] [Patent Literature]

[Patent Document 1] Japanese Patent Laid-Open No. Hei 11-86185

[Non-patent literature]

[Non-Patent Document 1] J Prokai, G Medioni, "3-D model based vehicle recognition," Applications of Computer Vision (WACV), 2009

[Non-Patent Document 2] A. Lai, G. Fung and N. Yung, "Vehicle type classification from visual-based dimension estimation" IEEE, Intelligent Transportation Systems, pages 201-206, 2001

Accordingly, it is an object of the present invention to provide a technique for calculating a specification (e.g., width, length, and height) of a vehicle by image processing that is robust to detection of feature points.

In one aspect of the invention, the vehicle specification measuring device includes an imaging device that sequentially captures a vehicle to acquire a frame image of the vehicle, and an image processing device that calculates a width of the vehicle based on the frame image. The image processing device includes a feature point detecting unit that detects a feature point of the first specification measurement, which is a feature point of the first surface of the rectangular parallelepiped that approximates the shape of the vehicle, and is located at the second of the rectangular parallelepiped in at least one of the frame images. The characteristic point of the surface is the feature point for measuring the second specification, and the feature point for measuring the third specification which is the feature point of the third surface of the rectangular parallelepiped; and the vehicle specification calculation means based on the feature point for the first specification measurement and the second The feature points for the measurement and the feature points for the third measurement are used to calculate the width of the vehicle. Here, the first surface is a rectangular parallelepiped surface corresponding to the first side surface of the vehicle, the second surface is a rectangular parallelepiped surface corresponding to the second side surface of the vehicle, and the third surface is a rectangular parallelepiped surface corresponding to the bottom surface of the vehicle. The vehicle specification calculation means calculates the amount of movement of the vehicle based on the movement amount of the characteristic point for measurement of the third specification, and uses the amount of movement of the vehicle to move the stereoscopic vision to the feature point for measurement of the first specification and the feature point for measurement of the second specification. Processing to calculate the feature points for the first specification measurement and the second The three-dimensional coordinates of the feature points for the measurement are measured, and the width of the vehicle is calculated based on the calculated three-dimensional coordinates of the first-order measurement feature point and the second-standard measurement feature point. The position of the first specification measurement feature point detected by the feature point detecting means has a degree of freedom in the plane of the first surface, and the position of the second specification measurement feature point detected by the feature point detecting means is based on There is freedom in the face of the 2 faces. Further, the position of the third-species measurement feature point detected by the feature point detecting means has a degree of freedom in the plane of the third surface.

In one embodiment, the feature point detecting mechanism may be configured to extract a moving area corresponding to the vehicle from the frame image, and extract a frame image corresponding to the line tangent to the vehicle in the real space for the moving area. The line is a tangent line, and the intersection point of the intersection of the extracted tangent points on the first surface is determined as the feature point for the first specification measurement. In this case, the feature point detecting means may be configured to determine the intersection point of the intersection of the extracted tangent points on the second surface as the feature point for the second specification measurement, or may be configured to be the third point in the intersection of the extracted tangent lines. The intersection point of the surface is determined as the feature point for the measurement of the third specification.

In one embodiment, the feature point detecting mechanism may be configured to extract a moving area corresponding to the vehicle from the frame image, and extract the frame image corresponding to the line tangent to the vehicle in the real space for the moving area. The line is a tangent line, and the feature point for the first specification measurement is determined by searching for the feature point along the first search path including the portion of the extracted tangent that is located on the first surface. In this case, the feature point detecting means may be configured to search the feature points along the second search path including the portion of the extracted tangent that is located on the second surface, thereby determining the feature points for the second specification measurement, and the feature points The detecting means may be configured to search for feature points along the third search path including the portion of the extracted tangent that is located on the third surface, thereby determining the feature points for the third specification measurement.

In one embodiment, the feature point detecting means may further include a first template database including a first template that is assumed to be present on the first side of the vehicle, and a feature point indicating the first template The first feature point position data of the position. In this case, the feature point detecting mechanism may also be configured as a frame image by using the first template database. When the template of the template is matched, the first matching position, which is the position of the first matching template, is obtained, and the feature point for the first specification measurement is present at the position determined based on the first matching position and the first feature point position data, and The feature points for the measurement of the first specification are detected.

Further, the feature point detecting means may further include a second template database including a second template which is assumed to be present on the second side surface of the vehicle, and a position indicating a position of the feature point in the second template 2 feature point location data. In this case, the feature point detecting means may be configured to obtain, by matching with the template of the second template of the second template database, the second matching position, which is the position of the most matching second template, for the frame image. The second specification measurement feature point is present at the position determined based on the second matching position and the second feature point position data, and the second specification measurement feature point is detected.

Furthermore, the feature point detecting means may further include a third template database including a third template which is assumed to be present on the bottom surface of the vehicle, and a third feature indicating the position of the feature point in the third template. Point location data. In this case, the feature point detecting means may be configured to obtain a third matching position, which is the position of the most matching third template, by matching the template of the third template of the third template database with respect to the frame image. The third specification measurement feature point is detected by the third specification measurement feature point at the position determined based on the third matching position and the third feature point position data.

In another aspect of the present invention, the vehicle specification measuring device includes an imaging device that sequentially captures a vehicle to acquire a frame image of the vehicle, and an image processing device that calculates a length of the vehicle based on the frame image. The image processing device includes a feature point detecting unit that detects a feature point of the third specification measurement, which is a feature point of the third surface of the rectangular parallelepiped that approximates the shape of the vehicle, and is located at the fourth of the rectangular parallelepiped for at least one of the frame images. The characteristic point of the surface is the characteristic point for measuring the fourth specification, and the characteristic point for measuring the fifth specification which is the characteristic point of the fifth surface of the rectangular parallelepiped; and the vehicle specification calculation means based on the third characteristic measurement feature point and the fourth The feature points for the measurement and the feature points for the fifth measurement are used to calculate the length of the vehicle. Here, the third surface is the surface of the rectangular parallelepiped corresponding to the bottom surface of the vehicle, and the fourth surface corresponds to the front surface of the vehicle. On the face of the cuboid, the fifth face is the face of the cuboid corresponding to the back of the vehicle. The vehicle specification calculation means calculates the amount of movement of the vehicle based on the amount of movement of the characteristic point for measurement of the third specification, and uses the amount of movement of the vehicle to perform moving stereoscopic vision on the fourth specification measurement feature point and the fifth specification measurement feature point ( The motion stereo) is used to calculate the three-dimensional coordinates of the fourth-element measurement feature point and the fifth-element measurement feature point, and based on the calculated third-dimensional measurement feature point and the fifth-dimensional measurement feature point three-dimensional coordinate, Calculate the length of the vehicle. The position of the third-species measurement feature point detected by the feature point detecting means has a degree of freedom in the plane of the third surface. The position of the fourth-element measurement feature point detected by the feature point detecting means has a degree of freedom in the plane of the fourth surface, and the position of the fifth-species measurement feature point detected by the feature point detecting means is based on the fifth There is freedom in the face of the face.

In still another aspect of the present invention, a vehicle specification measuring device includes an imaging device that sequentially captures a vehicle to acquire a frame image of the vehicle, and an image processing device that calculates a height of the vehicle based on the frame image. The image processing device includes: a feature point detecting unit that detects a feature point of the third specification measurement, which is a feature point of the third surface of the rectangular parallelepiped that approximates the shape of the vehicle, and is located at the cuboid in at least one of the frame images The characteristic point of the sixth surface is the characteristic point for measurement of the sixth specification; and the vehicle specification calculation means calculates the height of the vehicle based on the feature points for measurement of the third specification and the feature points for measurement of the sixth specification. Here, the third surface is a rectangular parallelepiped surface corresponding to the bottom surface of the vehicle, and the sixth surface is a rectangular parallelepiped surface corresponding to the top surface of the vehicle. The vehicle specification calculation means calculates the amount of movement of the vehicle based on the amount of movement of the feature points for the third measurement, and uses the amount of movement of the vehicle to perform stereoscopic stereoscopic processing on the feature points for measurement of the third specification and the feature points for measurement of the sixth specification. By calculating the three-dimensional coordinates of the third-element measurement feature point and the sixth-element measurement feature point, and calculating the height of the vehicle based on the calculated three-dimensional measurement feature point and the three-dimensional coordinate of the sixth-spec measurement feature point. . The position of the third-species measurement feature point detected by the feature point detecting means has a degree of freedom in the plane of the third surface, and the position of the sixth-species measurement feature point detected by the feature point detecting means is based on the sixth There is freedom in the face of the face.

In still another aspect of the present invention, the vehicle specification measuring method includes the steps of: acquiring a frame image of the vehicle by sequentially photographing the vehicle; and detecting, for the frame image, a feature point of the first surface of the rectangular parallelepiped that approximates the shape of the vehicle. 1st feature point for measurement measurement, a feature point for measurement of the second dimension on the second surface of the rectangular parallelepiped, and a feature point for measurement of the third specification which is a feature point on the third surface of the rectangular parallelepiped; and measurement based on the first specification The feature point, the second specification measurement feature point, and the third specification measurement feature point are used to calculate the width of the vehicle. Here, the first surface is a rectangular parallelepiped surface corresponding to the first side surface of the vehicle, the second surface is a rectangular parallelepiped surface corresponding to the second side surface of the vehicle, and the third surface is a rectangular parallelepiped surface corresponding to the bottom surface of the vehicle. The step of calculating the width of the vehicle includes the following steps: calculating the amount of movement of the vehicle based on the amount of movement of the characteristic point for measurement in the third specification; and measuring the characteristic point and the second specification for the first specification by using the amount of movement of the vehicle. The feature point is subjected to the moving stereoscopic processing to calculate the three-dimensional coordinates of the first-species measurement feature point and the second-element measurement feature point; and the three-dimensional coordinates of the first feature measurement feature point and the second specification measurement feature point. , calculate the width of the vehicle. The position of the first specification measurement feature point detected in the detection step has a degree of freedom in the plane of the first surface, and the position of the second specification measurement feature point detected in the detection step is based on the second surface. There is freedom in the plane. Further, the position of the third-species measurement feature point detected in the detection step has a degree of freedom in the plane of the third surface.

In still another aspect of the present invention, a vehicle specification measuring method includes the steps of: acquiring a frame image of a vehicle by sequentially photographing a vehicle; and detecting a third surface of a rectangular parallelepiped that is approximate to a shape of the vehicle for at least one of the frame images; The feature point is the feature point for measuring the third specification, the feature point for the fourth dimension of the rectangular parallelepiped, and the feature point for the fourth dimension of the rectangular parallelepiped, that is, the feature point of the fifth surface of the rectangular parallelepiped 5 The characteristic point for measurement; and the length of the vehicle is calculated based on the third specification measurement feature point, the fourth specification measurement feature point, and the fifth specification measurement feature point. Here, the third surface is a rectangular parallelepiped surface corresponding to the bottom surface of the vehicle, the fourth surface is a rectangular parallelepiped surface corresponding to the front surface of the vehicle, and the fifth surface is a rectangular parallelepiped surface corresponding to the back surface of the vehicle. The step of calculating the length of the vehicle includes the following steps: calculating the amount of movement of the vehicle based on the amount of movement of the characteristic point for measurement of the third specification; and measuring the characteristic point and the fifth specification for the fourth specification by using the amount of movement of the vehicle. The feature point is subjected to the moving stereoscopic processing, and the three-dimensional coordinates of the fourth-species measurement feature point and the fifth-species measurement feature point are calculated; and the third-dimensional measurement feature point and the fifth-specific measurement feature point are calculated based on the fourth feature. The coordinates are used to calculate the length of the vehicle. The position of the third-species measurement feature point detected in the detection step has a degree of freedom in the plane of the third surface. Further, the position of the fourth-species measurement feature point detected in the detection step has a degree of freedom in the plane of the fourth surface, and the position of the fifth-species measurement feature point detected in the detection step is based on the fifth There is freedom in the face of the face.

In still another aspect of the present invention, a vehicle specification measuring method includes the steps of: acquiring a frame image of a vehicle by sequentially photographing a vehicle; and detecting a third surface of a rectangular parallelepiped that is approximate to a shape of the vehicle for at least one of the frame images; The feature point is the feature point for the third measurement, and the feature point for the sixth measurement, which is the feature point on the sixth surface of the rectangular parallelepiped; and the feature point for the third measurement and the feature point for the sixth measurement, and the vehicle is calculated. The height. Here, the third surface is a rectangular parallelepiped surface corresponding to the bottom surface of the vehicle, and the sixth surface is a rectangular parallelepiped surface corresponding to the top surface of the vehicle. The step of calculating the height of the vehicle includes the following steps: calculating the amount of movement of the vehicle based on the amount of movement of the characteristic point for measurement in the third specification; and measuring the characteristic point and the sixth specification for the third specification by using the amount of movement of the vehicle The third-dimensional measurement feature point and the third-dimensional measurement feature point three-dimensional coordinates are calculated by using the feature point to perform the moving stereoscopic processing, and the third feature measurement feature point and the sixth specification measurement feature point are calculated. The three-dimensional coordinates are used to calculate the height of the vehicle. The position of the third-species measurement feature point detected in the detection step has a degree of freedom in the plane of the third surface, and the position of the sixth-species measurement feature point detected in the detection step is in the plane of the sixth surface. Have degrees of freedom.

The above vehicle specification measuring method can also be implemented by executing an program by an arithmetic device. In this case, the program can also be installed in the arithmetic unit using the recording medium on which the program is recorded.

According to the present invention, it is possible to provide a technique for calculating the specifications (e.g., width, length, and height) of a vehicle by image processing which is robust to the detection of feature points.

1‧‧‧ camera

2‧‧‧Image processing device

3‧‧‧Image Processing IC

4‧‧‧ External interface

5‧‧‧External memory device

6‧‧‧ memory

7‧‧‧ROM

7a‧‧‧Vehicle specification measurement processing program

10‧‧‧Vehicle specification measuring and processing device

11‧‧‧ Computing Module

12‧‧‧Image input interface

13‧‧‧Data input and output interface

14‧‧‧ memory controller

15‧‧‧ROM controller

16‧‧‧Internal busbar

21‧‧‧Image data

22‧‧‧Vehicle specifications

23‧‧‧ List of feature points for moving body detection

24‧‧‧ List of feature points for specification measurement

31, 32‧‧‧ side

33‧‧‧ bottom

34‧‧‧ positive

35‧‧‧Back

36‧‧‧Top

41‧‧‧Feature points

42a, 42b, 42c, 42d, 42e, 42f‧‧‧ feature points

43‧‧‧ vanishing point

44‧‧‧Processing area

51‧‧‧ intersection of tangent line 61 and tangent line 63

52‧‧‧ intersection of tangent lines 62 and 64

53‧‧‧ intersection of tangent line 61 and tangent 65

54‧‧‧ intersection of tangent line 64 and tangent 65

55‧‧‧ intersection of tangent line 63 and tangent 66

56‧‧‧ intersection of tangent line 62 and tangent 66

57‧‧‧ tangent point 67, 68

61‧‧‧tangential

62‧‧‧tangential

63‧‧‧ Tangential

64‧‧‧tangential

65‧‧‧tangential

66‧‧‧tangential

67‧‧‧ Tangential

68‧‧‧tangential

69‧‧‧tangential

71‧‧‧Template database

72‧‧‧Template database

73‧‧‧Template database

74‧‧‧Template database

75‧‧‧Template database

76‧‧‧Template database

D _M ‧‧‧Distance

Θ‧‧‧ angle

Fig. 1A is a conceptual diagram showing the position of a feature point for specification measurement for measuring the width of a vehicle in an embodiment of the present invention.

Fig. 1B is a conceptual diagram showing the position of a feature point for specification measurement for measuring the length of a vehicle in the present embodiment.

Fig. 1C is a conceptual diagram showing the position of a feature point for specification measurement for measuring the height of a vehicle in the present embodiment.

Fig. 2 is a block diagram showing an example of a hardware configuration of a vehicle specification measurement processing device in an embodiment.

Fig. 3A is a flow chart showing image processing for calculating the specifications of the vehicle in the embodiment.

Fig. 3B is a flow chart showing image processing for calculating the specifications of the vehicle in the embodiment.

Fig. 4 is a view conceptually showing the detection of the moving feature points in the embodiment.

Fig. 5 is a view conceptually showing a grouping of moving feature points in the embodiment.

Fig. 6 is a view conceptually showing the detection of the moving region in the embodiment.

Fig. 7A is a view showing a procedure for obtaining a tangent line and an intersection point for each moving region in the present embodiment.

Fig. 7B is a view showing a procedure for obtaining a tangent line and an intersection point for each moving region in the present embodiment.

FIG. 8A is a view showing a procedure of detecting characteristic points for measurement of the side surface and the bottom surface of the vehicle based on the tangent and the intersection obtained for each of the moving regions.

FIG. 8B is a view showing a procedure for detecting feature points for specification measurement on the front and back sides of the vehicle based on the tangent and the intersection obtained for each of the moving regions.

Fig. 8C is a view showing the procedure of detecting the feature points for specification measurement on the top surface of the vehicle based on the tangent and the intersection obtained for each of the moving regions in the present embodiment.

In the present embodiment, the feature points of the specification measurement feature points located on the side surface and the bottom surface of the vehicle are searched for by searching the feature points along the search route determined by the tangent and the intersection obtained for each of the movement regions. Figure.

In the present embodiment, the feature points are searched for along the search path determined by the tangent and the intersection obtained for each of the moving regions, and the order of the feature points for the specification measurement on the front and back sides of the vehicle is detected. Figure.

FIG. 9C is a view showing a sequence of detecting feature points on the top surface of the vehicle by searching for feature points along the search route determined by the tangent and the intersection obtained for each moving region in the present embodiment. .

Fig. 10 is a conceptual diagram showing the contents of a template database for detecting feature points for specification measurement on each surface of the vehicle in the embodiment.

Fig. 11 is a view conceptually showing a method of detecting feature points for specification measurement on each surface of a vehicle by template matching in the present embodiment.

Fig. 12 is a view conceptually showing the detection of entry and exit of the vehicle toward the processing area in the embodiment.

Fig. 13A is a view conceptually illustrating a moving stereoscopic vision process performed in the present embodiment.

Fig. 13B is a diagram conceptually illustrating a moving stereoscopic vision process performed in the present embodiment.

First, an outline of a vehicle specification measuring method according to an embodiment of the present invention will be described. In the present embodiment, xi images of the traveling vehicle are imaged sequentially (that is, images of the vehicle are captured at different times), and feature points are detected and tracked for the plurality of images. Further, a motion stereo processing is performed on the feature points. The mobile stereoscopic vision processing is a technique of capturing an image of a moving object at different times, and calculating a moving distance of the object by a method (which can also be calculated according to the captured image itself), and The three-dimensional coordinates of the object in the three-dimensional coordinate moving together with the object are calculated according to the moving distance of the captured image and the object. Even for images taken from the same place, the three-dimensional coordinates of the object can be calculated by a method similar to ordinary three-dimensional processing according to the position of the object only with different moving distances. In the present invention, the three-dimensional coordinates of the feature points (strictly speaking, the three-dimensional coordinates in the three-dimensional coordinates moving together with the vehicle) are calculated by the moving stereo vision processing, and the specifications of the vehicle are calculated based on the three-dimensional coordinates.

One of the measurement processes of the specifications of the vehicle in the present embodiment is a method for extracting feature points for calculation of the width, length, and height of the vehicle. 1A to 1C are views showing feature points for measurement of the width, length, and height of the vehicle, respectively. The present invention defines a rectangular parallelepiped 30 that approximates the shape of the vehicle, and uses a feature point located at a position along a specific face of the rectangular parallelepiped 30 as a feature point for measurement of the width, length, and height of the vehicle. Furthermore, the cuboid 30 is ultimately a virtual determiner. In FIGS. 1A to 1C, the symbols v ₁ to v ₈ represent the vertices of the rectangular parallelepiped 30. In addition, the following will be used for the measurement of the width, length, and height of the vehicle (in order to distinguish the feature point (the feature point for moving body detection) from the image detecting body as described below) Measuring feature points.

Specifically, as shown in FIG. 1A, in the calculation of the width of the vehicle, the following three types of feature points for measurement are used: feature point #1 for specification measurement: feature points located on the side 31 of the vehicle (strictly speaking, a feature point located at a position along a plane of the rectangular parallelepiped 30 corresponding to the side 31 of the vehicle)

Feature measurement feature point #2: a feature point located on the side 32 of the vehicle (strictly speaking, a feature point located at a position along the plane of the rectangular parallelepiped 30 corresponding to the side 32 of the vehicle

Feature measurement point #3: a feature point located on the bottom surface 33 of the vehicle (strictly speaking, a feature point located at a position along the surface of the rectangular parallelepiped 30 corresponding to the bottom surface 33 of the vehicle)

Here, the side surface 31 is the side surface of the two side surfaces of the vehicle opposite to the camera that measures the traveling vehicle, and the side surface 32 is the other side surface. The face of the rectangular parallelepiped 30 corresponding to the side face 31 has vertices v ₁ to v ₄ , and the face of the rectangular parallelepiped 30 corresponding to the side face 32 has vertices v ₅ to v ₆ .

In the present embodiment, if the feature measurement feature points #1, #2, and #3 satisfy the conditions of the side faces 31, 32, and the bottom face 33, respectively, it is allowed to extract from an arbitrary position. For example, for the side surface 31 opposed to the camera, a feature point located at an arbitrary position on the side surface 31 is allowed to be used as the feature measuring feature point #1.

Similarly, for the side surface 32 on the opposite side to the camera, the feature point located at any position on the side surface 32 is also allowed to be used as the feature point #2 for specification measurement. However, the side 32 is actually hidden from the vehicle, so in fact, the feature points extracted from any position of the contour of the vehicle along the side v ₅ v ₆ and the side v ₆ v ₇ are used as feature points on the side 32.

Further, with respect to the bottom surface 33, a feature point located at an arbitrary position on the bottom surface 33 is also allowed to be used as the feature measurement feature point #3. However, the bottom surface 33 is actually hidden from the vehicle, and therefore, in fact, feature points extracted from arbitrary positions of the contour of the vehicle along the side v ₃ v ₇ and the side v ₃ v ₄ are used as feature points on the bottom surface 33.

As described below, among the extracted feature measurement feature points #1 to #3, the specification measurement feature point #3 located on the bottom surface 33 is used to calculate the moving distance of the vehicle. Further, use The calculated moving distance is subjected to moving stereoscopic processing for the feature measurement feature points #1 and #2, and the three-dimensional coordinates of the specification measurement feature points #1 and #2 in the three-dimensional coordinate system moving together with the vehicle are calculated (for example, For the relative coordinates of the feature measurement feature point #3, the width of the vehicle is calculated based on the calculated three-dimensional coordinates of the feature measurement points #1 and #2.

On the other hand, as illustrated in FIG. 1B, in the calculation of the vehicle length, the following three types of feature points for measurement are used: feature point for measurement measurement #3: feature point at the bottom surface 33 of the vehicle

Feature measurement point #4: a feature point located on the front side 34 of the vehicle (strictly speaking, a feature point located at a position along the plane of the rectangular parallelepiped 30 corresponding to the front surface 34 of the vehicle)

Feature measurement point #5: a feature point located on the back side 35 of the vehicle (strictly speaking, a feature point located at a position along the plane of the rectangular parallelepiped 30 corresponding to the back surface 35 of the vehicle)

The feature measurement feature points #4 and #5 are also the same as the feature measurement feature points #1 to #3, and are allowed to be extracted from an arbitrary position if the conditions on the front surface 34 and the back surface 35 are satisfied. However, the position of the feature points extracted is actually different depending on the configuration of the camera. When the camera is configured to photograph a traveling vehicle from the rear, the back surface 35 becomes opposed to the camera. In this case, with respect to the back surface 35, the feature point located at any position on the back surface 35 is allowed to be used as the feature point #5 for specification measurement, and on the other hand, it is hidden from the front side of the vehicle 34. The feature point extracted at any position of the contour of the vehicle of v ₁ v ₅ and side v ₁ v ₄ is used as the feature point #4 for specification measurement on the front surface 34. The configuration of such a camera is illustrated in Figure 1B. On the other hand, when the camera is arranged in such a manner as to photograph the traveling vehicle from the front, the front surface 34 becomes opposed to the camera. In this case, for the front surface 34, the feature point located at any position on the front surface 34 is allowed to be used as the feature point #4 for specification measurement, and on the other hand, for the back surface 35 actually hidden on the vehicle, the self-slope can be The feature point extracted from any position of the outline of the vehicle on the side of the camera photograph is used as the feature point #5 for specification measurement on the back surface 35.

In the calculation of the length of the vehicle, the moving distance of the vehicle calculated based on the specification measurement feature point #3 located on the bottom surface 33 is used to perform the moving stereoscopic processing on the specification measurement feature points #4 and #5. Thereby, the three-dimensional coordinates of the feature measurement feature points #4 and #5 in the three-dimensional coordinate system moving together with the vehicle (for example, the relative coordinates of the feature measurement feature point #3) in the three-dimensional coordinate system are calculated, and are measured based on the calculated specifications. The length of the vehicle is calculated using the three-dimensional coordinates of feature points #4 and #5.

Further, as shown in FIG. 1C, the following two types of feature points for measurement are used for calculation of the height of the vehicle: feature point #3 for specification measurement: feature points located on the bottom surface 33 of the vehicle (strictly speaking, located along The feature point of the position of the face of the rectangular parallelepiped 30 corresponding to the bottom surface 33 of the vehicle)

Feature measurement feature point #6: a feature point located on the top surface 36 of the vehicle (strictly speaking, a feature point located at a position along the plane of the rectangular parallelepiped 30 corresponding to the top surface 36 of the vehicle)

Similarly to the feature measurement points #1 to #5, the specification measurement feature point #6 is allowed to be extracted from an arbitrary position if it satisfies the condition of the top surface 36. It should be noted that when photographing a vehicle traveling on a road, the camera is usually placed above the vehicle (for example, on a truss provided on a road), so that the top surface 36 of the vehicle is usually photographed as a whole.

In the calculation of the height of the vehicle, the moving distance of the vehicle calculated based on the specification measuring feature point #3 located on the bottom surface 33 is used to perform the moving stereoscopic processing on the specification measuring feature point #6. Thereby, the three-dimensional coordinates of the feature measurement feature points #3 and #6 in the three-dimensional coordinate system moving together with the vehicle (for example, the relative coordinates of the feature measurement feature point #3) in the three-dimensional coordinate system are calculated, and the measurement is performed based on the calculated specifications. The height of the vehicle is calculated using the three-dimensional coordinates of feature points #3 and #6.

Next, a vehicle specification measurement processing device and a vehicle specification measurement method according to an embodiment of the present invention will be described in detail.

Fig. 2 is a view showing a configuration example of the vehicle specification measurement processing device 10 of the present embodiment. Block diagram. The vehicle specification measurement processing device 10 includes a camera 1 and an image processing device 2. The camera 1 is a camera that captures a vehicle traveling on a road. The camera 1 generates image data 21 of an image (frame image) of the vehicle, and supplies the generated image data 21 to the image processing device 2. The image processing device 2 performs image processing on the image data 21 received from the camera 1 and measures the size of the vehicle captured in the image of the image data 21, more specifically, the width, length, and height of the vehicle. A vehicle specification data 22 indicating the width, length, and height of the vehicle is generated.

The image processing device 2 includes an image processing IC (integrated circuit) 3, an external interface 4, an external memory device 5, a memory 6, and a ROM (read only memory) 7. The external interface 4 is connected to the camera 1 and supplies the image data 21 received from the camera 1 to the image processing IC 3. The external memory device 5 stores data generated in image processing in the image processing device 2. The data stored in the external memory device 5 contains the vehicle specification data 22. The memory 6 is used as a work area for arithmetic processing of the image processing IC 3. The ROM 7 stores programs executed by the image processing IC 3. The program stored in the ROM 7 includes a vehicle specification measurement processing program 7a for executing image processing for measuring vehicle specifications.

Further, although the configuration of the image processing device 2 for calculating the vehicle standard from the image data 21 has been described above, the image processing device 2 may perform the vehicle type discrimination using the vehicle specification data 22. In this case, the program required for the vehicle type discrimination can also be installed in the ROM 7.

At this time, a rewritable non-volatile memory (for example, like a flash memory) can also be used as the ROM 7. In this case, the vehicle specification measurement processing program 7a (and, if necessary, the program for determining the vehicle type) may be externally mounted in the ROM 7. In the installation of the ROM 7 by the vehicle specification measurement processing program 7a, a recording medium that memorizes the vehicle specification measurement processing program 7a can also be used. In this case, a device for accessing the recording medium is provided in the image processing device 2. Further, an appropriate memory device (for example, a hard disk drive (HDD), a solid state drive) may be used instead of the ROM 7.

The image processing IC 3 includes a computing module 11, an image input interface 12, a data input/output interface 13, a memory controller 14, and a ROM controller 15. The arithmetic module 11, the image input interface 12, the data input/output interface 13, the memory controller 14, and the ROM controller 15 are connected by an internal bus bar 16. The calculation module 11 executes the vehicle specification measurement processing program 7a while using the memory 6 as a work area, and performs image processing for calculating the vehicle specification based on the image data 21. The image input interface 12 is an interface for receiving image data 21 from the camera 1. The data input/output interface 13 is an interface for accessing the external memory device 5. The memory controller 14 is an interface for performing access to the memory 6. Further, the ROM controller 15 is an interface for performing access to the ROM 7.

The image processing system for calculating the vehicle specifications described below is carried out using the hardware shown in FIG. 2. Hereinafter, image processing for calculating the vehicle specifications will be described in detail.

3A and 3B are flowcharts showing image processing for calculating a vehicle specification in an embodiment of the present invention. The vehicle specification measurement processing program 7a is a program encoding group for executing the image processing. The image processing system used to calculate the vehicle specifications is mainly divided into the following three stages:

1. Detection and tracking of feature points for moving body detection (steps S01~S04 of Figure 3A)

2. Detection and tracking of feature points for detection, separation, and specification measurement of moving objects (steps S05~S07 of Figure 3A)

3. Calculation of vehicle specifications based on feature points for specification measurement (steps S08 to S12 of Fig. 3B)

Hereinafter, each stage will be described in detail.

1. Detection and tracking of feature points for moving body detection

Referring to Fig. 3A, first, after the initialization of the image processing apparatus 2 is carried out (step S01), the image data 21 of the frame image obtained by the vehicle is taken in order (step S02). Further, the feature points included in the frame image shown in the image data 21 are detected, and each feature point is further tracked (step S03). However, the feature points that are detected and tracked in step S03 are not necessarily related to the above specifications. The measurement features are consistent. Therefore, in order to distinguish from the feature points for specification measurement, the feature points detected and tracked in step S03 are hereinafter referred to as feature points for moving body detection. In step S03, a list of feature points for moving body detection, which is a list of feature points for detecting the moving body to be detected and tracked, is generated. When the moving object detecting feature point is re-detected in the region where the moving body (in most cases, the vehicle) enters, the moving body detecting feature point is added to the moving body detecting feature point list 23. In addition, the moving object detecting feature point or the tracking failure moving feature detecting point that has disappeared outside the image is removed from the moving body detecting feature point list 23.

The acquisition of the image data 21 (step S02) and the detection and tracking of the feature points for detecting the moving body (step S03) are sequentially performed until the end signal is given to the image processing apparatus 2. When an end signal is given to the image processing apparatus 2 (step S04), the image processing in the image processing apparatus 2 ends.

2. Detection and tracking of feature points for detection, separation, and specification measurement of moving objects

At this stage, the feature points corresponding to the respective vehicles are detected from the feature points for moving body detection listed in the feature point list 23 for moving body detection, that is, the feature points for detecting the moving body currently being tracked. Specifically, first, feature points corresponding to the vehicle among the feature points for detecting a moving body listed in the moving object detecting feature point list 23 (that is, feature points other than the feature points not corresponding to the vehicle) are detected. As a moving feature point (step S05). More specifically, the feature points of the moving object detecting feature points listed in the moving object detecting feature point list 23 and the feature points that are moving but moving differently from the vehicle are removed. Since the vehicle traveling on the road moves substantially in the direction of the road, it can be determined that the feature point that does not perform such movement is a feature point that does not correspond to the vehicle.

Fig. 4 is a view showing an example of detection of a moving feature point in step S05. As shown in the left diagram of FIG. 4, in step S03, the feature points 41 are detected in a series of images, and further, the moving feature points 41 are tracked. In the left diagram of Figure 4, the feature points 41 being tracked are mutually The figure is connected by a line segment. The feature points judged to be stationary are removed from the moving object detection feature point list 23.

Further, the feature points satisfying the following conditions (1) to (3) among the tracked feature points are detected as the moving feature points corresponding to the vehicle:

(1) Tracking must occur throughout a continuous frame image of a specific number N _STD or more (a specific number N _STD is given as a parameter).

(2) The vector obtained by concatenating the starting point (the position at which the feature point is tracked on the image) and the current point (ie, the latest position on the image of the feature point) must be μ×N. Above _TRC . Here, μ is a specific parameter value (in terms of technology, a threshold value of the amount of movement per unit time), and N _TRC is the number of frame images for tracking the feature points.

(3) The angle between the vector from the starting point toward the current point and the vector from the starting point toward the vanishing point of the image must be below a certain threshold angle. The vanishing point of the image referred to herein refers to the vanishing point defined in the perspective view when the image of the vehicle is approximated to the perspective view. The vanishing point 43 is illustrated in the right diagram of FIG.

The right diagram of Fig. 4 shows an example of the moving feature points detected in step S05. The right diagram of Fig. 4 illustrates feature points 42a, 42b, 42c tracked over a particular number of consecutive frame images. Here, the symbol 42a-1 indicates the start point of the feature point 42a, and the symbol 42a-3 indicates the current point of the feature point 42a. The symbol 42a-2 indicates the position of the feature point 42a in the frame image between the frame image corresponding to the start point and the frame image corresponding to the current point. Similarly, the symbol 42b-1 indicates the start point of the feature point 42b, and the symbol 42b-3 indicates the current point of the feature point 42b. Symbol 42b-2 indicates the position of the feature point 42b in the frame image between the frame image corresponding to the start point and the frame image corresponding to the current point. Further, the symbol 42c-1 indicates the start point of the feature point 42c, and the symbol 42c-3 indicates the current point of the feature point 42c. Symbol 42c-2 indicates the position of the feature point 42c in the frame image between the frame image corresponding to the start point and the frame image corresponding to the current point.

For the feature point 42a, it is judged whether or not the size of the vector of the connection start point 42a-1 and the current point 42a-3 is μ×N _TR or more, and the vector from the start point 42a-1 toward the current point 42a-3 and the start point. Whether the angle between the points 42a-1 toward the vanishing point 43 is equal to or lower than the specific threshold value θ _TH1 . Based on the judgments, it is judged whether or not the feature point 42a is extracted as the moving feature point.

The same judgment is made for the feature points 42b and 42c. Hereinafter, in step S05, feature points 42a, 42b, and 42c are detected as moving feature points.

As shown in FIG. 3A, after the detection of the moving feature points (step S05), the grouping of the detected moving feature points is performed. (Step S06). Referring to Fig. 5, the grouping is performed based on the distance from each feature point to the vanishing point 43 and the direction of the vector from the starting point toward the current point. Furthermore, as the distance from the feature point to the vanishing point 43, the distance from the current point to the vanishing point 43 (illustrated by the distance D _M in FIG. 5) may be used, and the distance from the starting point to the vanishing point 43 may be used. Moreover, the average distance during the tracking period of the feature points can also be used. By this grouping, the feature points corresponding to one group of one vehicle are classified into the same cluster.

Fig. 5 shows an example of the grouping in step S06. Referring to Fig. 5, six feature points are extracted in step S05: feature points 42a, 42b, 42c, 42d, 42e, and 42f. In this case, for each of the feature points 42a, 42b, 42c, 42d, 42e, and 42f, the distance to the vanishing point 43 and the direction of the vector from the starting point toward the current point are calculated. In FIG. 5, the distance from each feature point to the vanishing point 43 is defined as the distance D _M from the current point to the vanishing point 43, and the direction of the vector from the starting point toward the current point is defined as the horizontal and horizontal directions of the image. The angle of formation.

For example, for the feature point 42a, the distance from the current point 42a-3 to the vanishing point 43 and the angle θ from the start point 42a-1 toward the current point 42a-3 and the horizontal direction of the screen are calculated. Similarly to the other feature points 42b, 42c, 42d, 42e, and 42f, the distance from the current point to the vanishing point 43 and the direction from the start point to the current point vector are calculated. The grouping of the feature points 42a, 42b, 42c, 42d, 42e, 42f is performed based on the calculated distance from the current point to the vanishing point 43 and the direction of the vector from the starting point toward the current point. In the example of Figure 5 As a result of the grouping, the feature points 42a, 42b, and 42c are classified into clusters corresponding to the first vehicle, and the feature points 42d, 42e, and 42f are classified into clusters corresponding to the second vehicle.

As shown in FIG. 3A, after the grouping of the moving feature points, the moving regions corresponding to the respective vehicles are extracted from the image, and the feature points for the specification measurement are detected and tracked for each of the moving regions (step S07).

First, as shown in FIG. 6, from the moving feature points of the respective clusters, the moving region in which the vehicle is photographed in each image is extracted. The mobile area is extracted for each cluster. The moving area is extracted as an area of the image corresponding to the shape of the polygon of the vehicle outline. At this time, the moving area is extracted such that the moving feature points of the respective clusters become the vertices of the polygon.

Further, for the moving area corresponding to each vehicle, a process of detecting the feature points for the specification measurement and a process of tracking the feature points for the specification measurement are performed. Here, as described with reference to FIG. 1A, the feature measurement feature points #1 and #2 are respectively located at the feature points of the side faces 31 and 32 of the vehicle, and the specification measurement feature point #3 is located at the feature point of the bottom surface 33 of the vehicle. Further, the feature measurement feature point #4 is located at the feature point of the front surface 34 of the vehicle, and the specification measurement feature point #5 is located at the feature point of the rear surface 35 of the vehicle. Further, the feature measurement feature point #6 is located at a feature point of the top surface 36 of the vehicle. Note that the number of feature points #1 for the measurement of the extracted specifications is not necessarily limited to one, and a plurality of feature points #1 for specification measurement may be extracted. The same applies to the feature points #2 to #6 for other specifications.

Here, the tracking of the feature points for the specification measurement may be performed by detecting the feature points for the specification measurement for each frame image, and associating the feature points for the specification measurement between the frame images. Further, the tracking of the feature points for the specification measurement can be performed by detecting the feature points for the specification measurement on the frame image (initial frame image) in which the moving region is initially present, and in the subsequent frame image, Dynamic tracking is performed on the feature points for the specification measurement.

In either case, at least the processing of the feature points for the initial frame detection specification measurement is performed. Therefore, first, the processing of the feature points for the measurement specification measurement will be described. The processing for detecting characteristic points for measurement measurement is performed by at least one of the following methods (1) to (3). Here, please note that the number of feature points #1 for the specification measurement that is extracted is not necessarily limited to one, and a plurality of feature points #1 for specification measurement may be extracted. Therefore, a plurality of feature points #1 for specification measurement can be extracted by a plurality of methods (1) to (3) below. The same applies to the feature points #2 to #6 for other specifications.

method 1):

The method of detecting the feature points for measurement measurement (1) extracts a tangent tangent to each vehicle, and finds the intersection of the tangent intersections. The feature points for specification measurement are selected and determined from among the intersections obtained. The term "tangent line" as used herein refers to a line corresponding to an image of a tangent to a vehicle that is tangent to the vehicle in real space. 7A and 7B show an example of processing for finding the intersection of the tangent line and the like which are tangent to the vehicle. There is also a case where the tangent line is tangent to the moving area in the image (the tangent shown by the symbols 61 to 66), but is not necessarily limited thereto (for example, the tangent shown by the symbols 67 to 69).

As shown in the column (1) of FIGS. 7A and 7B, in the method (1), first, the tangent lines 61, 62 corresponding to the lines on the road surface and parallel to the lanes in the actual space are extracted. In the actual space, the angle at which the line on the road surface parallel to the lane is photographed in the image is determined by the spatial configuration of the camera 1 and is known in advance. Using this fact, a tangent corresponding to the line on the road surface and parallel to the lane is extracted. Here, the tangent 61 is a tangent to the near side of the vehicle (taken from the camera 1) which is extracted from a line parallel to the line of the lane, and the tangent 62 is located at a tangent to the distal side of the vehicle.

Further, as shown in the column (2) of FIGS. 7A and 7B, a tangent line corresponding to a line on the road surface and perpendicular to the lane line in the actual space is extracted. Similarly to the line parallel to the lane in the actual space, at what angle the image is perpendicular to the line of the vehicle in the actual space, which is determined by the spatial arrangement of the camera 1, is known in advance. Use the matter Really, extract the tangent corresponding to the line perpendicular to the lane in the actual space. In the column (2) of FIG. 7A and FIG. 7B, the tangent line 63 is a tangent line which is extracted from the line corresponding to the line perpendicular to the lane, and is located at the front side of the vehicle (when viewed from the camera 1), and the tangent line 64 is located in the vehicle. The tangent to the distal side.

In addition to this, the intersection 51 between the tangent line 61 and the tangent line 63 is obtained, and the intersection 52 of the tangent lines 62 and 64 is obtained.

Further, as shown in the column (3) of FIG. 7A and FIG. 7B, a tangent corresponding to a line perpendicular to the road surface in the actual space is extracted. At what angle in the image the line perpendicular to the road surface in the actual space is photographed in advance, so that the tangent to the moving area corresponding to the line perpendicular to the lane in the actual space can be extracted. In the column (3) of FIG. 7A and FIG. 7B, the tangent line on the distal end side of the vehicle which is extracted from the line corresponding to the line perpendicular to the road surface is indicated by the symbol 65, and is located at the tangent line on the near side of the vehicle. It is represented by symbols 66 and 67. Here, the tangent 67 is extracted as a line passing through the intersection 51 of the tangent lines 61, 63 located on the near side of the vehicle.

In addition, as shown in the column (3), the intersection point of the tangent 61 and the tangent 65, the intersection 54 of the tangent 64 and the tangent 65, the intersection 55 of the tangent 63 and the tangent 66, and the intersection of the tangent 62 and the tangent 66 are obtained. 56.

Further, as shown in the column (3) of FIG. 7A and FIG. 7B, a tangent 68, 69 corresponding to a line which is located farther from the road surface in the actual space and which is in a plane parallel to the road surface and which is perpendicular or parallel to the lane line is extracted. . Here, the tangent 68 is a tangent corresponding to the line perpendicular to the lane, and the tangent 69 is a tangent corresponding to the line parallel to the lane. Tangent 68 is extracted as a line through intersection 56. In addition to this, the intersection 57 of the tangent lines 67, 68 is obtained, and the tangent line 69 is extracted as a line passing through the intersection point 57.

8A to 8C are views showing a method of determining feature points for measurement of each specification from among the obtained intersection points. Specifically, FIG. 8A shows the feature measuring points #1 and #2 located on the side faces 31 and 32 of the vehicle, and the feature measuring feature points located on the bottom surface 33 of the vehicle. Diagram of #3's decision method. In addition, FIG. 8B is a view showing a method of determining the feature measuring point #5 located on the front surface 34 of the vehicle and the specification measuring feature point #5 located on the back surface 35 of the vehicle. Fig. 8C is a view showing a method of determining the feature point #6 for specification measurement on the top surface 36 of the vehicle. Further, the method of determining the feature measurement feature point #3 located on the bottom surface 33 of the vehicle is not only shown in FIG. 8A but also shown in FIGS. 8B and 8C.

Referring to Fig. 8A, feature point #1 for specification measurement is selected from the intersections of the side faces 31 of the vehicle (i.e., intersections 51, 53, 54, 57) at the intersections obtained as described above. Here, all of the intersection points 51, 53, 54, and 57 can be determined as the feature measurement feature point #1, and one or a plurality of intersection points of the self-intersection points 51, 53, 54, and 57 can be appropriately determined as Feature point #1 for specification measurement.

Further, the feature measurement feature point #2 is selected from the intersections of the side faces 32 of the vehicle (that is, the intersection points 52, 55, 56). Here, all of the intersection points 52, 55, and 56 can be determined as the feature measurement feature point #2, and one or more of the intersection points 52, 55, and 56 can be appropriately determined as the specification measurement feature. Point #2.

Further, the feature measurement feature point #3 is selected from the intersections of the bottom faces 33 of the vehicle (that is, the intersection points 51, 53, 55). Here, all of the intersection points 51, 53, and 55 can be determined as the feature measurement feature point #3, and one or a plurality of intersection points which are appropriately selected from the self-intersection points 51, 53, and 55 can be determined as the feature measurement characteristics. Point #3.

Further, as illustrated in FIG. 8B, the feature measurement feature point #4 is selected from among the intersections of the front faces 34 of the vehicle (that is, the intersection points 52, 53, 54). Here, all of the intersection points 52, 53, and 54 can be determined as the feature measurement feature point #4, and one or more of the intersection points 52, 53, 54 can be appropriately determined as the specification measurement feature. Point #4.

Further, the feature measurement feature point #5 is selected from the intersections of the back faces 35 of the vehicle (that is, the intersection points 51, 55, 56, 57). Here, all of the intersection points 51, 55, 56, and 57 can be determined as the feature measurement feature point #5, and the self-intersection points 51, 55, 56, and 57 can also be included. One or more of the appropriate selection points is determined as the feature point #5 for specification measurement.

Further, as illustrated in FIG. 8C, the feature measurement feature point #6 is selected from among the intersections (ie, intersections 52, 56, 57) located at the top surface 36 of the vehicle. Here, all of the intersection points 52, 56, and 57 can be determined as the feature measurement feature point #6, and one or more of the intersection points 52, 56, and 57 can be appropriately determined as the specification measurement feature. Point #6.

Method (2):

The method (2) for detecting the feature points for measurement measurement is obtained by searching for feature points along the search path determined by the obtained tangent and intersection points, and then searching for feature points in the same manner as the method (1). Search for feature points that are easier to track. The feature points obtained by the search are determined as feature points for specification measurement. 9A to 9C are views showing a method of determining feature points for measurement of each specification in the method (2).

Referring to Fig. 9A, in the determination of the feature measurement feature point #1 located on the side surface 31, the path including the portion located on the side surface 31 among the tangent lines obtained as described above is determined as the search path, and the periphery of the search path is performed. Search for feature points. At this time, the intersection point is used to obtain a portion of the tangent located on the side 31 when determining the search path. More specifically, the search path including the portion connecting the intersection 54 and the intersection 53 in the tangent 65, the portion of the intersection 53 and the intersection 51 in the tangent 62, and the portion of the tangent 67 connecting the intersection 51 and the intersection 57 Search for feature points around. The feature point obtained by the search is determined as the feature measurement feature point #1. The number of feature points #1 for specification measurement may be one or plural.

In the determination of the feature measurement point #2 located on the side surface 32, a path including a portion of the obtained tangent located on the side surface 32 is determined as a search path, and a feature point is searched for the periphery of the search path. More specifically, the feature points are searched for the periphery of the search path including the portion of the tangent 66 and the intersection 56 and the portion of the tangent 62 and the intersection 52. The feature points obtained by the search are determined as Feature point #2 for specification measurement. The number of feature points #2 for specification measurement may be one or plural.

Further, in the determination of the feature measurement point #3 located on the bottom surface 33, a path including a portion of the obtained tangent located on the bottom surface 33 is determined as a search path, and feature points are formed around the search path. search for. More specifically, the feature points are searched for the periphery of the search path including the portion of the tangent 61 and the intersection 51 and the portion of the tangent 63 connecting the intersection 51 and the intersection 55. The feature point obtained by the search is determined as the feature measurement feature point #2. The number of feature points #2 for specification measurement may be one or plural.

Further, referring to FIG. 9B, in the determination of the feature measurement feature point #4 located on the front surface 34, the path including the portion located on the front surface 34 among the tangent lines obtained as described above is determined as the search path, and the search path is determined. Search for feature points around. More specifically, the feature points are searched for the periphery of the search path including the portion of the tangent 65 connecting the intersection point 53 and the intersection 54 and the portion of the tangent 64 connecting the intersection 54 and the intersection 52. The feature point obtained by the search is determined as the feature measurement feature point #4. The number of feature points #4 for specification measurement may be one or plural.

Further, in the determination of the feature measurement point #5 located on the back surface 35, a path including a portion of the obtained tangent located on the back surface 35 is determined as a search path, and feature points are searched for the periphery of the search path. . More specifically, the portion including the intersection 55 and the intersection 56 in the tangent 66, the portion of the intersection 56 and the intersection 57 in the tangent 68, the portion of the intersection 57 and the intersection 51 in the tangent 67, and the tangent 63 are included. The feature points are searched for around the search path of the portion of the intersection 51 and the intersection 55. The feature point obtained by the search is determined as the feature measurement feature point #5. The number of feature points #5 for specification measurement may be one or plural.

Further, referring to FIG. 9C, in the determination of the feature measurement feature point #6 located on the top surface 36, the path including the portion of the obtained tangent located on the top surface 36 is determined as the search path. The path searches for feature points around the search path. More specifically, the feature points are searched for the periphery of the search path including the portion of the tangent 68 and the intersection 56 and the portion of the tangent 62 and the intersection 52. The feature point obtained by the search is determined as the feature measurement feature point #6. The number of feature points #6 for specification measurement may be one or plural.

Method (3):

The method (3) for detecting the feature points for measurement measurement performs a feature point search for each surface of the vehicle directly (that is, without referring to the tangent and the intersection obtained by the above method (1)). The feature points obtained by the search are determined as feature points for specification measurement. 10 to 11 are views showing a method of determining feature points for specification measurement in the method (3).

In the case of the method (3) of detecting the feature points for measuring the specification, a library of templates constituting the parts existing on each side of the vehicle is prepared, and template matching using the template is performed. Fig. 10 is a view conceptually showing an example of contents of a database of prepared templates. The database of prepared templates contains the following six databases:

‧ Template database 71 containing templates for the parts on the side 31 of the camera side

‧ Template library 72 containing templates for parts located on the side 32 opposite the camera

‧ Template database 73 containing templates for the parts on the bottom surface 33

‧ Template database 74 containing templates for the parts on the front side 34

‧Template database 75 containing templates for the parts on the back 35

‧ Template database 76 containing templates for parts on top surface 36

In each of the template databases 71 to 76, feature point position data indicating the position of the feature points in the template is attached to the template of each part.

For example, for the template database 71, a template created by an image of a part (such as a tire, a window, a rear view mirror, and a vehicle lamp) assumed to be located on the side surface 31 is registered, and the position of the feature point is designated for each template. The black dots in Fig. 10 indicate the feature points assigned to the respective templates. In order to correspond A plurality of types of vehicles are provided, and in the template database 71, a plurality of templates are prepared for each part. Similarly to the other template databases 72 to 76, templates created by images assumed to be located on the respective faces are registered, and feature points are assigned to the respective templates. The template databases 71-76 are stored in an appropriate memory mechanism, such as the external memory device 5 illustrated in FIG.

Figure 11 is a diagram conceptually illustrating the detection of template matching feature points for specification measurement. In addition, FIG. 11 illustrates the detection of the feature measurement feature point #5 located on the back surface 35 using the template registered in the template database 75. The feature measurement feature point #5 located on the back surface 35 performs matching processing with the template included in the template database 75 (i.e., the template of the part located on the back surface 35) for the entire frame image of the processing target. For each position of each frame image, the similarity is obtained for each template, and the template that best matches the position of the template and the position (matching position) of the template are calculated for each part. Further, the position of the feature point for specification measurement is determined based on the position of the template to be matched most and the position of the feature point specified by the feature point position data for the template. The feature points for specification measurement are detected by this sequence.

Although the detection of the feature point #5 for the specification measurement on the back surface 35 has been described above, the template matching is performed in the same manner for the other surfaces (the side surface 31, 32, the bottom surface 33, the front surface 34, and the top surface 36). Detection of feature points for measurement.

The detection of the feature points for the specification measurement of the above methods (1) to (3) can also be performed for each frame image. In this case, the tracking of the feature points for the specification measurement is also performed by the method of associating the feature points for the specification measurement between the frame images.

The detection of the feature points for the specification measurement of the above methods (1) to (3) can also be performed on the frame image (initial frame image) in which the moving region first appears. In this case, for the subsequent frame image, the dynamic tracking using the appropriate algorithm for the detected feature points for the specification measurement is performed. The detection and tracking of the feature points for the specification measurement in step S07 are performed in the above manner.

Referring again to FIG. 3A, the detection and tracking of the feature points for the specification measurement in step S07 are performed, and the history data of the feature points for the measurement of the specification and the tracking are displayed. The data of the position of the feature measurement feature points in the frame image inscribed is recorded in the specification measurement feature point list 24. The specification measurement feature point list 24 is recorded in an appropriate memory mechanism (for example, the external memory device 5 of FIG. 2). The specification measurement feature point list 24 is generated for each movement area whenever a moving area appears in the image.

3. Calculation of the specifications of the vehicle based on the feature points for measurement

After the detection and tracking processing of the feature points for the specification measurement, the specification of the vehicle is calculated using the feature points for the measurement. More specifically, as shown in FIG. 3B, first, a time point at which each vehicle as a whole enters a specific processing area specified in each frame image is detected (step S08), and further, the vehicle is detected to be withdrawn from the processing area. Time point (step S09). Figure 12 is a conceptual representation of the detection of vehicle entry and exit of a processing area. A specific region of each frame image (for example, a region defined by a lower portion of the frame image) is defined as the processing region 44, and a time point at which the entire moving region corresponding to each vehicle enters the processing region 44 is detected, and the self-processing region 44 is specified. The point in time for the exit.

Further, in the history data of the feature measurement feature points of each of the moving regions (that is, each vehicle) recorded in the specification measurement feature point list 24, the time point and the self-processing area in which the entire vehicle enters the processing region 44 as a whole The time history-related history data between the time points of the exits is 44, and the vehicle specification data 22 indicating the specifications of the vehicle is calculated (step S13). The vehicle specification data 22 is calculated by the specification measurement feature point in the frame image at each time between the time point when each vehicle enters the processing area 44 and the time point when the vehicle is exited from the processing area 44. The location data is processed by moving stereoscopic processing. The three-dimensional coordinates of the feature points for specification measurement at each time (strictly speaking, the three-dimensional coordinates in the three-dimensional coordinates moving together with the vehicle) are calculated by moving the stereoscopic vision processing, and the specifications of the vehicle are obtained based on the three-dimensional coordinates.

Figs. 13A and 13B are conceptual views for explaining the moving stereoscopic vision processing for obtaining the three-dimensional coordinates of the feature points for measurement of each specification in the present embodiment. Referring to Fig. 13A, in the moving stereoscopic vision processing of the present embodiment, the amount of movement d _C of the vehicle in the actual space is calculated based on the amount of movement in the frame image of the feature measuring feature point #3 located on the bottom surface 33 of the vehicle. The correspondence between the amount of movement in the frame image of the point on the actual space road surface and the amount of movement in the actual space of the point can be obtained in advance based on the spatial configuration of the camera 1. By using this fact, the feature point for specification measurement can be calculated from the amount of movement in the frame image of the feature measurement point #3 (which is approximated to the point on the road surface in the actual space) of the bottom surface 33 of the vehicle. The amount of movement d _C in the actual space of #3.

As shown in FIG. 13B, the moving stereoscopic vision processing is performed based on the specification measurement feature points #1, ~#6 before the calculated movement amount d _{C is} moved. In other words, the three-dimensional coordinates of the feature measurement feature points #1 to #6 are calculated by the moving stereoscopic processing performed by the specification measurement feature points #1 to #6 by the movement amount d _C movement.

When the three-dimensional coordinates of the feature points for measurement of each specification are calculated, the three-dimensional coordinates are used to calculate the width, length, and height of the vehicle, which is the specification of the vehicle. The width of the vehicle is calculated based on the three-dimensional coordinates of the feature measuring points #1 and #2 located on the side faces 31 and 32 as shown in FIG. 1A. Further, as shown in FIG. 1B, the length of the vehicle is calculated based on the three-dimensional coordinates of the feature measuring points #4 and #5 located on the front surface 34 and the back surface 35. Further, as shown in FIG. 1C, the height of the vehicle is calculated based on the three-dimensional coordinates of the feature measuring points #3 and #6 located on the bottom surface 33 and the top surface 36.

Further, as shown in FIG. 3B, vehicle specification data 22 indicating the specifications of the calculated vehicle is generated. The vehicle specification data 22 is stored in the external memory device 5. Thereafter, the calculated specification of the vehicle is outputted using an appropriate output device (for example, a display device) (step S11), and the specification measurement feature point list 24 corresponding to the vehicle whose specification is calculated is deleted (step S12). As described above, the process for calculating the specifications of the vehicle is completed.

As described above, in the present embodiment, feature points for measurement measurement are extracted for each surface of the vehicle, and the feature point for measurement measurement is subjected to moving stereoscopic processing, thereby calculating the three-dimensional coordinates of the feature points for measurement of each specification. . The specifications (width, length, and height) of the vehicle are calculated based on the calculated three-dimensional coordinates of the feature points for each measurement. Here, this is The measurement processing of the vehicle specification in the embodiment has a certain degree of freedom for the position of the feature measurement feature points extracted on each side of the vehicle, whereby the feature points for specification measurement can be detected by the image processing of the toughness. As described above, according to the measurement processing of the vehicle standard in the present embodiment, the specifications (for example, the width, the length, and the height) of the vehicle can be calculated by the tough image processing.

Although the embodiments of the present invention have been specifically described above, the present invention is not limited to the above embodiments. The Applicant is aware that the invention can be implemented with various modifications. For example, although the embodiment in which the width, the length, and the height of the vehicle are all calculated is described above, the entire width, length, and height of the vehicle may not be calculated. The present invention is applicable to the case of calculating at least one of the width, the length, and the height of the vehicle. In this case, only the necessary ones of the specification measurement feature points #1 to #6 can be detected.

1‧‧‧ camera

2‧‧‧Image processing device

3‧‧‧Image Processing IC

4‧‧‧ External interface

5‧‧‧External memory device

6‧‧‧ memory

7‧‧‧ROM

7a‧‧‧Vehicle specification measurement processing program

10‧‧‧Vehicle specification measuring and processing device

11‧‧‧ Computing Module

12‧‧‧Image input interface

13‧‧‧Data input and output interface

14‧‧‧ memory controller

15‧‧‧ROM controller

16‧‧‧Internal busbar

21‧‧‧Image data

22‧‧‧Vehicle specifications

Claims (20)

A vehicle specification measuring device comprising: an imaging device that sequentially captures a vehicle to obtain a frame image of the vehicle; and an image processing device that calculates a width of the vehicle based on the frame image; the image processing device includes The feature point detecting means detects a feature point of the first specification measurement which is a feature point of the first surface of the rectangular parallelepiped which approximates the shape of the vehicle, and is located at the second surface of the rectangular parallelepiped in at least one of the frame images. The feature point of the second specification measurement, and the feature point of the third specification measurement which is the feature point of the third surface of the rectangular parallelepiped; and the vehicle specification calculation means based on the first specification measurement feature point and the above (2) the width of the vehicle is calculated by the feature point for measurement of the specification and the feature point for measurement of the third specification; the first surface is a surface of the rectangular parallelepiped corresponding to the first side surface of the vehicle, and the second surface is the vehicle of the vehicle a surface of the rectangular parallelepiped corresponding to the second side surface, wherein the third surface is a surface of the rectangular parallelepiped corresponding to a bottom surface of the vehicle, and the vehicle specification The exiting mechanism calculates the amount of movement of the vehicle based on the amount of movement of the feature point for the measurement of the third specification, and performs the feature point for the first specification measurement and the feature point for the second specification measurement using the amount of movement of the vehicle. By moving the stereoscopic vision processing, the three-dimensional coordinates of the feature point for the first specification measurement and the feature point for the second specification measurement are calculated, and the first feature measurement feature point and the second specification measurement feature point are calculated based on the calculated Calculating the width of the vehicle based on the three-dimensional coordinates; and the feature points for measuring the first specification detected by the feature point detecting means The position has a degree of freedom in the surface of the first surface, and the position of the second specification measurement feature point detected by the feature point detecting means has a degree of freedom in the surface of the second surface, and the feature point is The position of the third-species measurement feature point detected by the detecting means has a degree of freedom in the surface of the third surface. The vehicle specification measuring device according to claim 1, wherein the feature point detecting mechanism is configured to extract a moving region corresponding to the vehicle from the frame image, and extract the tangent to the vehicle in the actual space. The line corresponding to the line in the frame image is a tangent line, and the intersection of the extracted intersections of the tangent lines at the first surface is determined as the feature point for the first specification measurement. The vehicle specification measuring device according to claim 2, wherein the feature point detecting means is configured to determine, as the second specification measuring feature point, an intersection point of the intersection of the extracted tangent lines at the second surface. The vehicle specification measuring device according to claim 2 or 3, wherein the feature point detecting means is configured to determine, as the third specification measuring feature point, an intersection point of the intersection of the extracted tangent lines at the third surface. The vehicle specification measuring device according to claim 1, wherein the feature point detecting mechanism is configured to extract a moving area corresponding to the vehicle from the frame image, and extract the tangent to the vehicle in the actual space for the moving area The line corresponding to the line in the frame image, that is, the tangent line, searches for the feature point along the first search path including the portion of the extracted tangent that is located on the first surface, thereby determining the first measurement. Feature points. The vehicle specification measuring device according to claim 5, wherein the feature point detecting means is configured to search for a feature point along a second search path including a portion of the extracted tangent that is located on the second surface, thereby The feature points for the measurement of the second specification described above are determined. The vehicle specification measuring device according to claim 5 or 6, wherein the feature point detecting means is configured to search for a feature point along a third search path including a portion of the extracted tangent that is located on the third surface, thereby The feature points for the measurement of the third specification described above are determined. The vehicle specification measuring device according to claim 1, wherein the feature point detecting means includes a first template database including a first template which is assumed to be present on a first side surface of the vehicle, and the first template a first feature point position data of a position of a feature point in the template, wherein the feature point detecting means is configured to obtain a maximum for the frame image by matching with a template of the template of the first template database The first matching position, which is the position of the first template, is matched, and the first characteristic measuring point is present at the position determined based on the first matching position and the first feature point position data, and the first feature is detected. Feature points for specification measurement. The vehicle specification measuring device according to claim 1 or 8, wherein the feature point detecting means includes a second template database including a second template of a part assumed to exist on a second side surface of the vehicle, and a representation The second feature point position data of the position of the feature point in the second template, wherein the feature point detecting mechanism is configured to match the template of the second template of the second template database with respect to the frame image The second matching position, which is the position of the second template that matches the most matching, is obtained, and the second characteristic measuring feature point exists in the position determined based on the second matching position and the second feature point position data. The feature points for measuring the second specification described above are detected. The vehicle specification measuring device according to claim 8 or 9, wherein the feature point detecting means has a third template database including a third template of a part assumed to exist on a bottom surface of the vehicle, and indicating the above The third feature point position data of the position of the feature point in the third template, wherein the feature point detecting mechanism is configured to match the template of the third template of the third template database with the template of the third template database The third matching position, which is the position of the most matching third template, is obtained, and the third characteristic measuring feature point exists in the position determined based on the third matching position and the third feature point position data, and is detected. The feature points for the measurement of the third specification described above. The vehicle specification measuring device according to claim 1, wherein the feature point detecting means detects, for the frame image, a feature point for measuring the fourth dimension located on the fourth surface of the rectangular parallelepiped, and a fifth feature measuring point located in the rectangular parallelepiped The feature point of the surface is the characteristic point for measuring the fifth gauge, the fourth surface is a surface of the rectangular parallelepiped corresponding to the front surface of the vehicle, and the fifth surface is a surface of the rectangular parallelepiped corresponding to the back surface of the vehicle, and the vehicle specification The calculation means performs the moving stereoscopic processing on the fourth specification measurement feature point and the fifth specification measurement feature point by using the movement amount of the vehicle, and calculates the fourth specification measurement feature point and the fifth specification. Calculating the three-dimensional coordinates of the feature points for measurement, and calculating the length of the vehicle based on the calculated fourth-dimensional measurement feature points and the three-dimensional coordinates of the fifth-species measurement feature points; and the above-described detection by the feature point detecting means The position of the feature point for the fourth specification measurement has a degree of freedom in the plane of the fourth surface, and is detected by the feature point detecting mechanism 5 position of the first system specification of measuring points wherein a degree of freedom in the plane of the first surface 5. The vehicle specification measuring device according to claim 1 or 11, wherein the feature point detecting means detects a feature point of the sixth specification that is a feature point located on the sixth surface of the rectangular parallelepiped for the frame image, and the sixth surface system The face of the rectangular parallelepiped corresponding to the top surface of the vehicle, The vehicle specification calculation means performs a moving stereoscopic process on the third specification measurement feature point and the sixth specification measurement feature point using the movement amount of the vehicle, thereby calculating the third specification measurement feature point and the 6 three-dimensional coordinates of feature points for measurement, and calculating the height of the vehicle based on the calculated third-dimensional measurement feature points and the three-dimensional coordinates of the sixth-species measurement feature points; detected by the feature point detecting mechanism The position of the feature point for measuring the sixth specification is a degree of freedom in the surface of the sixth surface. A vehicle specification measuring device comprising: an imaging device that sequentially captures a vehicle to obtain a frame image of the vehicle; and an image processing device that calculates a length of the vehicle based on the frame image; the image processing device includes The characteristic point detecting means detects, for at least one of the frame images, a characteristic point of the third specification measurement which is a feature point of the third surface of the rectangular parallelepiped which is approximately the shape of the vehicle, and a feature of the fourth surface of the rectangular parallelepiped The feature point for measuring the fourth specification and the feature point for measuring the fifth specification which is the feature point of the fifth surface of the rectangular parallelepiped; and the vehicle specification calculation means based on the feature point for measuring the third specification and the fourth The characteristic measuring point and the fifth characteristic measuring feature point are used to calculate the length of the vehicle; the third surface is a surface of the rectangular parallelepiped corresponding to the bottom surface of the vehicle, and the fourth surface corresponds to the front surface of the vehicle. a surface of the rectangular parallelepiped, wherein the fifth surface is a surface of the rectangular parallelepiped corresponding to a back surface of the vehicle, and the vehicle specification calculation mechanism The displacement of the vehicle of the first 3 Specification Measurement calculating an amount of movement of the feature point according to, and the amount of movement of the vehicle, on the upper The fourth-standard measurement feature point and the fifth-element measurement feature point are subjected to a moving stereoscopic vision process, thereby calculating the three-dimensional coordinate of the fourth-species measurement feature point and the fifth-element measurement feature point, and Calculating the length of the vehicle by calculating the three-dimensional coordinates of the fourth characteristic measurement feature point and the fifth specification measurement feature point; and the position of the third specification measurement feature point detected by the feature point detection means is The surface of the third surface has a degree of freedom, and the position of the fourth-element measurement feature point detected by the feature point detecting means has a degree of freedom in the surface of the fourth surface, and the feature point detecting means The position of the characteristic point for measuring the fifth specification described above is determined to have a degree of freedom in the surface of the fifth surface. A vehicle specification measuring device comprising: an imaging device that sequentially captures a vehicle to obtain a frame image of the vehicle; and an image processing device that calculates a height of the vehicle based on the frame image; the image processing device includes The feature point detecting means detects, for at least one of the frame images, a feature point of the third specification measurement which is a feature point of the third surface of the rectangular parallelepiped which is approximately the shape of the vehicle, and a sixth surface of the rectangular parallelepiped The feature point is a feature point for measuring the sixth specification; and the vehicle specification calculation means calculates the height of the vehicle based on the third specification measurement feature point and the sixth specification measurement feature point; and the third surface is The surface of the rectangular parallelepiped corresponding to the bottom surface of the vehicle, wherein the sixth surface is the surface of the rectangular parallelepiped corresponding to the top surface of the vehicle, and the vehicle specification calculation means calculates the vehicle based on the movement amount of the characteristic measurement point of the third specification. The amount of movement, using the amount of movement of the above vehicle, for the above The third dimension measurement feature point and the sixth specification measurement feature point are subjected to a moving stereoscopic vision process, thereby calculating the three-dimensional coordinates of the third specification measurement feature point and the sixth specification measurement feature point, and calculating the three-dimensional coordinates based on the calculated Calculating the height of the vehicle by the feature point of the third specification measurement and the three-dimensional coordinates of the feature point for measuring the sixth specification, and the position of the third specification measurement feature point detected by the feature point detection means is The plane of the third surface has a degree of freedom, and the position of the characteristic point for measuring the sixth specification detected by the feature point detecting means has a degree of freedom in the plane of the sixth surface. A vehicle specification measuring method includes the steps of: acquiring a frame image of the vehicle by sequentially capturing a vehicle; and detecting, by the frame image, a feature point of a first surface of a rectangular parallelepiped that approximates a shape of the vehicle, that is, a first specification measurement a characteristic point, a feature point for measuring the second specification that is a feature point of the second surface of the rectangular parallelepiped, and a feature point for measuring the third specification that is a feature point of the third surface of the rectangular parallelepiped; and the measurement based on the first specification Calculating the width of the vehicle using the feature point, the second characteristic measurement feature point, and the third specification measurement feature point; the first surface is a surface of the rectangular parallelepiped corresponding to the first side surface of the vehicle, and the second surface The surface of the rectangular parallelepiped corresponding to the second side surface of the vehicle, the third surface of the rectangular parallelepiped surface corresponding to the bottom surface of the vehicle, and the step of calculating the width of the vehicle includes the following steps: measuring according to the third specification Calculating the amount of movement of the vehicle by using the amount of movement of the feature point; and using the amount of movement of the vehicle, the feature points for measuring the first specification and The second specification measurement feature point is subjected to a moving stereoscopic vision process to calculate a three-dimensional coordinate of the first specification measurement feature point and the second specification measurement feature point; Calculating the width of the vehicle based on the calculated first-order measurement feature point and the third-dimensional measurement feature point three-dimensional coordinates; and the position of the first-standard measurement feature point detected in the detection step is The surface of the first surface has a degree of freedom, and the position of the second specification measurement feature point detected in the detecting step has a degree of freedom in the surface of the second surface, and the first detected in the detecting step 3 The position of the feature point for measurement of the specification has a degree of freedom in the plane of the third surface. A vehicle specification measuring method includes the steps of: acquiring a frame image of the vehicle by sequentially capturing a vehicle; and detecting, at least one of the frame images, a feature point of a third surface of a rectangular parallelepiped that approximates a shape of the vehicle The feature point for the third measurement, the feature point for the fourth specification measurement, which is the feature point of the fourth surface of the rectangular parallelepiped, and the feature point for the fifth specification measurement which is the feature point of the fifth surface of the rectangular parallelepiped; The third gauge measurement feature point, the fourth specification measurement feature point, and the fifth specification measurement feature point are used to calculate the length of the vehicle; the third surface is a surface of the rectangular parallelepiped corresponding to the bottom surface of the vehicle, The fourth surface is a surface of the rectangular parallelepiped corresponding to the front surface of the vehicle, and the step of calculating the length of the vehicle on the surface of the rectangular parallelepiped corresponding to the back surface of the vehicle on the fifth surface includes the following step: measuring according to the third specification Calculating the amount of movement of the vehicle by the amount of movement of the feature point; and measuring the fourth specification using the amount of movement of the vehicle Performing a moving stereoscopic process on the feature point of the fifth specification measurement, and calculating the three-dimensional coordinates of the fourth-spec measurement feature point and the fifth-spec measurement feature point; and the calculated fourth specification Measurement feature points and the above-mentioned fifth specification measurement Calculating the length of the vehicle by using the three-dimensional coordinates of the feature points; the position of the third-species measurement feature point detected in the detecting step has a degree of freedom in the surface of the third surface, and is detected in the detecting step. The position of the characteristic point for measuring the fourth specification has a degree of freedom in the surface of the fourth surface, and the position of the characteristic point for measuring the fifth specification detected in the detecting step is in the plane of the fifth surface. Degree of freedom. A vehicle specification measuring method includes the steps of: acquiring a frame image of the vehicle by sequentially capturing a vehicle; and detecting, at least one of the frame images, a feature point of a third surface of a rectangular parallelepiped that approximates a shape of the vehicle The feature point for the third measurement, the feature point for the sixth measurement that is the feature point of the sixth surface of the rectangular parallelepiped; and the feature point for the measurement of the third specification and the feature point for the measurement of the sixth specification, the vehicle is calculated. The height of the third surface is the surface of the rectangular parallelepiped corresponding to the bottom surface of the vehicle, and the sixth surface is the surface of the rectangular parallelepiped corresponding to the top surface of the vehicle, and the step of calculating the height of the vehicle includes the following steps: The amount of movement of the vehicle is calculated based on the amount of movement of the feature point for the measurement of the third specification, and the third feature measurement feature point and the sixth specification measurement feature point are subjected to moving stereoscopic processing using the movement amount of the vehicle. And calculating the three-dimensional coordinates of the third feature measurement feature point and the sixth specification measurement feature point; Calculating the height of the vehicle by calculating the three-dimensional coordinate of the third feature measurement feature point and the sixth specification measurement feature point; and the position of the third specification measurement feature point detected in the detection step is There is freedom in the face of the third face, The position of the characteristic point for measuring the sixth specification detected in the detecting step is a degree of freedom in the plane of the sixth surface. A recording medium storing a program, and when executing the program, the program causes the computing device to perform the following steps: detecting a frame image of the vehicle obtained by sequentially photographing the vehicle, and detecting the vehicle located in the approximate vehicle The feature point of the first surface of the rectangular parallelepiped, that is, the feature point for the first specification measurement, the feature point for the second specification measurement, which is the feature point of the second surface of the rectangular parallelepiped, and the feature point of the third surface of the rectangular parallelepiped The feature point for the measurement of the third specification; and the feature point of the first specification measurement, the feature point for the second measurement, and the feature point for the measurement of the third specification, the width of the vehicle is calculated; the first surface is a surface of the rectangular parallelepiped corresponding to the first side surface of the vehicle, wherein the second surface is a surface of the rectangular parallelepiped corresponding to the second side surface of the vehicle, and the third surface is a surface of the rectangular parallelepiped corresponding to a bottom surface of the vehicle; The step of width of the vehicle includes the following steps: calculating the amount of movement of the vehicle according to the amount of movement of the feature points for measuring the third specification; using the vehicle The moving amount is subjected to a moving stereoscopic process for the feature point for measurement of the first specification and the feature point for measurement of the second specification, thereby calculating the three-dimensional coordinates of the feature point for the first specification measurement and the feature point for the second specification measurement And calculating the width of the vehicle based on the calculated first-order measurement feature point and the third-dimensional measurement feature point three-dimensional coordinates; and detecting the position of the first-standard measurement feature point detected in the detecting step a degree of freedom in a surface of the first surface, and a position of the second characteristic measurement feature point detected in the detecting step is a degree of freedom in a surface of the second surface. The position of the third-species measurement feature point detected in the above-described detection step has a degree of freedom in the surface of the third surface. A recording medium storing a program, and when executing the program, the program causes the computing device to perform the steps of: sequentially capturing a vehicle to obtain a frame image of the vehicle; and for at least one of the frame images, The feature point for measuring the third dimension, which is the feature point of the third surface of the rectangular parallelepiped which is approximately the shape of the vehicle, and the feature point for measuring the fourth specification, which is the feature point of the fourth surface of the rectangular parallelepiped, and the first feature of the rectangular parallelepiped The feature point of the fifth dimension is the fifth feature measurement feature point; and the length of the vehicle is calculated based on the third feature measurement feature point, the fourth specification measurement feature point, and the fifth specification measurement feature point; The third surface is a surface of the rectangular parallelepiped corresponding to a bottom surface of the vehicle, the fourth surface is a surface of the rectangular parallelepiped corresponding to a front surface of the vehicle, and the fifth surface is a surface of the rectangular parallelepiped corresponding to a back surface of the vehicle. The step of calculating the length of the vehicle includes the step of calculating the movement of the vehicle based on the amount of movement of the feature points for measuring the third specification The fourth specification measurement feature point and the fifth specification measurement point are calculated by performing a moving stereoscopic process on the fourth specification measurement feature point and the fifth specification measurement feature point using the movement amount of the vehicle. Using the three-dimensional coordinates of the feature points; and calculating the length of the vehicle based on the calculated fourth-dimensional measurement feature points and the three-dimensional coordinates of the fifth-species measurement feature points; and the third specification detected in the detecting step The position of the measurement feature point has a degree of freedom in the surface of the third surface, and the position of the fourth specification measurement feature point detected in the detection step has a degree of freedom in the surface of the fourth surface. The position of the fifth-species measurement feature point detected in the above-described detection step has a degree of freedom in the surface of the fifth surface. A recording medium storing a program, and when executing the program, the program causes the computing device to perform the steps of: sequentially capturing a vehicle to obtain a frame image of the vehicle; and for at least one of the frame images, The feature point for measuring the third dimension that is the third surface of the rectangular parallelepiped that is approximately the shape of the vehicle, and the feature point for the sixth specification that is the feature point of the sixth surface of the rectangular parallelepiped; Calculating a height of the vehicle by a feature point for measurement measurement and a feature point for measurement of the sixth specification; the third surface is a surface of the rectangular parallelepiped corresponding to a bottom surface of the vehicle, and the sixth surface corresponds to a top surface of the vehicle The step of calculating the length of the vehicle on the surface of the rectangular parallelepiped includes the step of calculating the amount of movement of the vehicle based on the amount of movement of the characteristic point for measuring the third gauge, and measuring the third gauge using the amount of movement of the vehicle. The third stereoscopic measurement is calculated by performing the moving stereoscopic processing using the feature points and the feature points for measuring the sixth specification. Point and the three-dimensional coordinate of the feature point for measuring the sixth specification; and calculating the height of the vehicle based on the calculated third-dimensional measurement feature point and the three-dimensional coordinate of the sixth-spec measurement feature point; The position of the third feature measuring feature point detected has a degree of freedom in the surface of the third surface, and the position of the sixth specification measuring feature point detected in the detecting step is the sixth surface. There is freedom in the plane.