JP2002190023A - Device and method for discriminating car model, and storage medium storing car model discriminating program readable in computer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device and a method for discriminating a car model, and a storage medium storing the method allowing real-time processing and having less restriction on the installation conditions of a camera. SOLUTION: This device for discriminating the car model has the camera 1 and electronic equipment 2. The electronic equipment 2 comprises an image signal processor 4, a model storage device 5, a model projection processor 6 and a discriminating device 7. The image signal processor 4 extracts a vehicle region T including a vehicle, from an image picked up by the camera 1. The model storage device 5 stores every vehicle model as a three-dimensional shape model. The model projection processor 6 reads the vehicle model from the model storage device 5 and converts the read vehicle model into an image viewed from picked-up position in the picked-up image. The discriminating device 7 obtains the area of the converted vehicle model, the area of the vehicle region T and the intersecting area of these areas and discriminates the car model (a large car, a standard-sized car or a two-wheeler). The reduction of a calculation cost and the like can be attained by using the three-dimensional shape model as the vehicle model.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle type discriminating apparatus, a vehicle type discriminating method, and a computer-readable storage medium storing a program for performing vehicle type discrimination.

[0002]

2. Description of the Related Art Traffic volume for each type of vehicle (large-sized car, ordinary car, small-sized car, two-wheeled vehicle, etc.) passing on a general road or an expressway,
A method of performing wide-area monitoring of the speed of each type of vehicle traveling on the road, etc., is to determine the type of the vehicle using an image captured by a visible camera and obtain desired information (traffic volume, speed, etc.),
There were several methods described below.

[0003] Methods for determining the type of vehicle include, for example,
A method of reading the model number on the license plate of the vehicle, a method of focusing on dimensions such as the height and width of the vehicle, a method of using the positions and intervals of headlights and tail lamps of the vehicle, a boundary between the vehicle surface and the vehicle and the background A method of extracting an edge component present in a part, a method using the size of a vehicle region in an image captured by a visible camera, and the like have been proposed.・ Method using license plate In this method, an imaging camera is arranged at a position where the license plate of the vehicle can be imaged, the license plate is extracted from the captured image by edge processing, and the license plate is classified according to the extracted license plate. The number is detected by template matching to determine the vehicle type. In this method, the vehicle type can be determined almost certainly, but there are problems such as a missing license plate detection and a long processing time for searching for a vehicle (vehicle type) number. Therefore, it is not suitable for wide area monitoring and real-time processing. -Method using the size of the vehicle This method extracts a vehicle from the captured image by edge processing, and calculates the vehicle width, vehicle length, and vehicle length from the extracted area.
Calculate and calculate dimensions such as vehicle height. The obtained numerical value is compared with information stored in the database in advance to determine the vehicle type. This method has a low computational cost for discrimination and is effective for real-time and wide-area monitoring.However, depending on the dimensions to be measured, there are installation conditions such as the camera mounting position and the number of cameras. Did not work. In addition, the area of the image to be imaged must be set so that one vehicle is imaged. Therefore, if the imaged area is an intersection or a congested road, a plurality of vehicles are densely packed and one vehicle is imaged. The dimensions of the vehicle could not be obtained.・ Method using headlights and tail lamps In this method, the positions of headlights and tail lamps are determined by edge processing, and the spacing and mounting position of headlights and the like are determined from the determined positions of headlights and tail lamps. And the vehicle type from the position. in this way,
It is effective at night and in places where lights and lamps are used in tunnels.However, in the daytime, the position of headlights and tail lamps must be determined accurately, which requires computational costs and is necessary for real-time and wide-area monitoring. Is not suitable. Also, when the road surface is wet in rainy weather, the light is reflected on the road surface,
It was difficult to determine the true headlight and tail lamp positions. Method Using Edge Component This method extracts a silhouette of a windshield, a side glass, a hood, and a background of a vehicle existing in a vehicle, and determines a vehicle type from the shape of each of the extracted components. This method is disclosed in, for example, Japanese Patent Application Laid-Open No. 11-353581, in which a silhouette of a vehicle is cut out by differential processing, and this silhouette area is further divided into parts such as a headlight, a hood, a windshield, a roof, and a rear part after the roof. By comparing each of the divided parts with data stored in advance, highly accurate vehicle type determination has been achieved. However, in this method, each part is divided when a vehicle in the captured image has an edge irrelevant to each part, or when a plurality of vehicles are detected overlapping with each other due to a camera installation situation. I couldn't. -Method using the size of the vehicle area This method extracts the area that is likely to be a vehicle by edge processing from the captured image, and determines the vehicle type by comparing it with the size of the data that is held in advance for each vehicle type I do. In this method, the calculation cost is low and real-time processing is possible.However, even if the detected area is not a vehicle, the size (area) of the detected area is a predetermined area set for each vehicle type. However, since this area is determined as a vehicle, the reliability of the determination is low.

[0004] As a method of increasing reliability, in a method described in Japanese Patent No. 2953232, a vehicle passing through an area imaged by a camera is imaged in advance and stored as image data. The determination is made by comparing the vehicle with the vehicle in the actually captured image. However, when the installation position and the angle of view of the camera are changed, it is necessary to create a plurality of image data again.

[0005]

As described above, in the conventional vehicle type discriminating apparatus, real-time processing could not be performed, camera installation conditions were restricted, and discrimination could not be performed when a plurality of vehicles overlapped. It was difficult to determine the desired vehicle type because of the difficulty in extracting each part of the vehicle.

Accordingly, the present invention has been made in view of the above-mentioned conventional problems, and is capable of real-time processing, has less restrictions on camera installation conditions, and can determine a desired vehicle type. It is another object of the present invention to provide a computer-readable storage medium storing a vehicle type determining method and a program for performing vehicle type determination.

[0007]

In order to achieve the above object, a vehicle type discriminating apparatus according to the present invention comprises: an image pickup means for picking up an image in an area where a moving object passes; Moving body region extracting means for extracting a region including the moving body as a moving body region; storage means for storing a three-dimensional shape of a plurality of kinds of moving bodies as a moving body model for each vehicle type of the moving body; A moving body model image converting means for converting the moving body model image into a moving body model image converted into a view of the imaging direction of the moving body on the image, and a case where the moving body model image is superimposed on the moving body on the image. Vehicle type discriminating means for comparing the degree of overlap and discriminating the vehicle type of the moving body from the comparison result.

[0008] In addition, according to the vehicle type discriminating method of the present invention, an image of an area in which a moving object passes is taken as an imaging range, and an area including the moving object is extracted from the taken image as a moving object area. And storing the three-dimensional shapes of the plurality of types of moving objects as a moving object model for each vehicle type of the moving object, and converting the moving object model into a view of the imaging direction of the moving object on the image. A moving body model image converting step of converting the moving body model image into a moving body model image, and a step of comparing the degree of overlap when the moving body model image is superimposed on the image, and determining a vehicle type of the moving body from the comparison result. Have.

A storage medium according to the present invention is a storage medium in which a program for determining the type of a vehicle is stored so as to be readable by a computer. And extracting a region including the moving object from the captured image as a moving object region, storing a three-dimensional shape of a plurality of types of moving objects as a moving object model for each vehicle type of the moving object, To
It is converted into a moving body model image converted to the appearance of the imaging direction of the moving body on the image, and the degree of overlap when the moving body model image is superimposed on the image is compared.From this comparison result, It is a computer-readable storage medium storing a program for determining a vehicle type of a moving object.

[0010] The program of the present invention has a function of causing a computer to capture an image having an area where a moving object passes as an imaging range, and an area including the moving object from the captured image as a moving object area. A function of extracting, a function of storing a three-dimensional shape of a plurality of types of moving objects as a moving object model for each vehicle type of the moving object, and a method of displaying the moving object model in the appearance of the imaging direction of the moving object on the image. A function for converting the moving object model image into a converted one, and a function for comparing the degree of overlap when the moving object model image is superimposed on the image and discriminating the vehicle type of the moving object from the comparison result. Is done.

According to this configuration, the vehicle model has a three-dimensional shape model of the vehicle type to be determined, and the size of the image when the vehicle model is displayed on the image captured by the image capturing means and the vehicle area And the size of the area where the image and the vehicle area intersect, the type of the vehicle can be determined, so that the calculation cost can be suppressed and the restriction on the installation location of the device can be reduced. In addition, when a vehicle to be determined is extracted, even if a vehicle running in parallel with the vehicle is extracted at the same time, the vehicle type can be determined substantially correctly. In addition, since the determination process per vehicle is very simple, the vehicle type can be determined without increasing the calculation amount even when a plurality of detection regions are provided or on a road with a large traffic volume. In addition, it is possible to simultaneously calculate the vehicle tracking, the detection of the pedestrian, the speed of a vehicle group (in a state where a plurality of vehicles are substantially densely packed) while determining the vehicle, and the like.

[0012]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 to FIG. 15 show a first embodiment.

FIG. 1 is a perspective view when the first embodiment is arranged on a road.

The first embodiment is directed to a camera 1 serving as an image pickup means.
And an electronic device 2 such as a computer. The electronic device 2 is provided in the camera 1, but may be provided in a separate housing from the camera 1 main body.

The camera 1 is located on a road through which a moving vehicle passes.
The road is fixed to a column 3 disposed on the side of (lane L1, L2) so that the road is imaged from above. The height at which the camera 1 is fixed is, for example, 6 to 8 m, and an area imaged by the camera 1 is a detection area D. This detection area D is 2 in FIG.
Although it is a lane and an area of approximately 8 × 8 m square, it may be one lane or three lanes or more, and can be arbitrarily set by a user. Further, the camera 1 is arranged in the traveling direction (upstream) of a vehicle passing through a road on which the camera 1 is installed, and captures an image in front of the vehicle. Conversely, it may be arranged downstream and image the rear of the vehicle. The grayscale value of the image captured by the camera 1 is, for example, 256 gradations, and this gradation can be set arbitrarily.

FIG. 2 is a block diagram of the first embodiment. The electronic device 2 includes an image signal processing device 4 and a model storage device.
5. It has a model projection processing device 6 (model image conversion means) and a discrimination device 7.

An image signal of an image picked up by the camera 1 is sent to an image signal processing unit 3. The image signal is transmitted to the image signal processing unit 3 by wire if the camera 1 and the electronic device 2 are connected, and wirelessly transmitted if not connected.
When an image signal is received wirelessly, a transmitting / receiving device such as an antenna, a filter, and an amplifier is separately provided.

The image signal processing section 4 extracts a region including the vehicle region T from the image (detection region D) captured by the camera 1. The vehicle region T refers to a region (comprising a plurality of pixels) at least including a vehicle on a captured image. The extraction of the vehicle region T uses, for example, a method such as a background difference or an inter-frame difference. The vehicle extraction region T may include a shadow of the vehicle due to the weather.

The model storage device 5 stores the three-dimensional shape of a vehicle for each vehicle type (large-sized vehicle, ordinary vehicle, small-sized vehicle, two-wheeled vehicle, etc.) as a vehicle model. This vehicle model is, for example, CAD
It has a data structure like data and expresses the features of the three-dimensional shape of the vehicle by a plurality of points (coordinates). For example,
In the case of an ordinary car, the number of points is 16 points, and 4 points
4 points, 4 hoods, 4 roofs, 4 trunks,
Expresses the characteristics of the vehicle.

The model projection processing device 6 converts an arbitrary vehicle model into a vehicle model image. The vehicle model image is data represented by a two-dimensional image by connecting a plurality of points representing the vehicle model. The lane width in which the vehicle travels and the traveling direction of the vehicle in the detection area D are stored in advance as data. This vehicle model image shows the lane width in the detection area D.
This is an image obtained by converting the (real length), the traveling direction of the vehicle, and the extracted vehicle region and converting the vehicle model arranged at the position of the extracted vehicle region T into a captured image (view).

The discriminating device 7 determines the size (the number of pixels) of the vehicle area T,
The vehicle type is determined based on the size of the vehicle model image and the like.
Specifically, the degree of overlap in the case where the vehicle region T and the vehicle model image overlap is determined.

The vehicle type discriminating apparatus includes a storage device such as an MO and an HDD for storing the discrimination result, a display device for displaying the discrimination result and the wide area monitoring result, and a printer for outputting the discrimination result and the wide area monitoring result. It is also possible to separately provide equipment.

FIG. 4 is a block diagram of the electronic apparatus shown in FIG.
This will be described with reference to the flowchart of FIG.

As shown in FIG. 3, the electronic device 2 has at least a CPU 10 for performing various types of arithmetic processing and a memory 11 for storing the calculation results and vehicle models, and displays the calculation results and the like as peripheral devices of the electronic device 2. Display 12, a printer 13 for outputting calculation results, etc., and a mouse 14 for operating the electronic device 2.
And a keyboard 15. The electronic device 2 as shown in FIG. 3 is not a case where the electronic device 2 is arranged in the camera 1, but a case where the electronic device 2 is arranged in a management center or the like which centrally processes and manages image signals from a plurality of cameras 1.

(1) An image of the detection area D is captured by the camera 1 (S1). The captured image is transmitted to the electronic device 2 by wire or wirelessly, and stored in the memory 11. When storing in the memory 11, the imaging time (storage time) when the image was captured (stored)
Is stored in a format that also includes The camera 1 always captures the detection area D.

(2) Noise processing is performed on the captured image, and a background difference method (or an inter-frame difference method) is used.
Extract an area consisting of a plurality of pixels whose gray value has changed (S
2). Then, the size of the extracted region is obtained. Most of the moving objects on the road are vehicles such as automobiles, and it is unlikely that humans or other objects exist. Therefore, the extracted area is regarded as the vehicle area T. The process of extracting the vehicle region T from the image is performed by the CPU 10, and the vehicle region T is stored in the memory 11 together with the size of the vehicle region T.

Note that, along with the extraction of the vehicle area T, the lane width on the image at the end of each lane in the vehicle area T in the traveling direction (FIG.
(W1 and W2 in 5) are obtained in pixel units, and stored in the memory 11 in association with the vehicle region T.

(3) The CPU 10 checks whether or not the vehicle area T is included in a specific area in the detection area D (S3). The specific region is a region indicated by oblique lines in the perspective view of the detection region D indicating the specific region in FIG. 6, and is a rectangular region having a width of approximately 2 meters in the traveling direction. The vehicle area T falls within this specific area.
If the front end (in the running direction) is included, the rear end (in the running direction) is within the detection area D, and almost one vehicle is imaged. When the vehicle region T does not fall within the specific region, the process proceeds to S1 because there is no vehicle in the detection region D or the vehicle cannot be extracted. If the vehicle area T is within the specific area, the process proceeds to the next step. The specific area can be set by the user as an arbitrary area in the captured image.

(4) When a vehicle exists in the detection area D, a vehicle model of a large vehicle is read out of the memory 11 among the vehicle models stored for each vehicle type, and the read vehicle model and image From the width of the upper lane, the traveling direction of the vehicle, and the vehicle area T, an image is created which is converted into a view (view in the imaging direction) in the vehicle area T of the vehicle model of the large vehicle. Then, using the image of the converted vehicle model as the model region M, the area (number of pixels) of the model region M is obtained (S
Four). The conversion of the vehicle model and the area of the model area M are calculated and obtained by the CPU 10, and the model area M and the area are stored in the memory 11. Note that the vehicle model to be read first may be another vehicle instead of a large vehicle.

(5) Using the model area M as a mask pattern, the model area M is moved within the vehicle area T, and the maximum value of the intersection area C where the model area M and the vehicle area T intersect is obtained (S5).
The movement of the model area M and the calculation of the intersection area C are performed by the CPU 10, and the maximum value of the intersection area C is stored in the memory 11 in association with the vehicle model. Vehicle area T, model area M and intersection area C
Is shown in FIG.

(6) The CPU 10 checks whether or not the intersection area C has been obtained for all vehicle models (S6). Initially, only the intersection area C for a large vehicle is determined, so the process proceeds to the next step. The intersection area C for all vehicle models
If is required, proceed to (9).

(7) If the intersection area C has not been obtained for all vehicle models, the other vehicle models,
For example, a vehicle model of an ordinary vehicle is read, and a model area M and an area of the model area M of the vehicle model of the ordinary vehicle are determined (S7). The determined model area M and the area of the model area M are stored in the memory 11 in association with the vehicle model.

(8) The maximum value of the intersection area C between the model area M and the vehicle area T obtained in (7) is obtained (S8). The obtained maximum value of the intersection area C is stored in the memory 11 in association with the vehicle model. Proceed to (S6).

(9) If the vehicle region T is an arbitrary model region n (however,
The criterion of the accuracy (n) that is (n = 1, 2, 3, 4) is obtained and stored in the memory 11 (S9). This accuracy (n) is defined as C / M.

The accuracy (n) is such that when the selected vehicle model is correct for the vehicle region T, the size of the intersection area C approaches the size of the model region M,

(Equation 1) Similarly, if the selected model area M is smaller than the vehicle area T,

(Equation 2) On the other hand, if the selected model area M is larger than the vehicle area T,

(Equation 3) (10) T / M is obtained for each vehicle model (S10). Asked
T / M is stored in the memory 11 in association with the vehicle model, and this calculation is performed by the CPU 10.

(11) A new accuracy (n) is obtained for each vehicle model based on a predetermined relationship and stored (S11). The predetermined relationship is the accuracy if T / M obtained in (S10) is T / M <1
(n) = Accuracy (n) × M / T is used, and if T / M ≧ 1, it means using the equation of accuracy (n) = Accuracy (n) × T / M.

The calculation of the new accuracy (n) is performed by the CPU 10, and the new accuracy (n) is stored in the memory 11.

(12) The vehicle type is determined using the accuracy (n) (S1).
2). The discrimination of the vehicle type is performed in the vehicle region where the vehicle model having the maximum value of the accuracy (n) obtained in (S11) is captured by the camera 1.
It is determined that the vehicle is included in T. Accuracy from memory 11
(n) is read, the maximum value is calculated, and the operation to determine the vehicle type is C
Performed by PU10. The determined vehicle type is stored in the memory 11 in association with the vehicle area T together with the imaging time.

By the above steps (1) to (12),
The camera 1 determines the type of the vehicle in the captured image.

The determined vehicle type can be output to the display 12 and the printer 13 as appropriate. In addition, by performing such determination for a plurality of vehicles or installing a plurality of cameras, it is possible to monitor a wide area such as a traffic volume and a speed of a passing vehicle.

Next, when the model area M is larger than a two-wheeled vehicle such as a large vehicle, if the vehicle area T is larger in proportion to the model area M, the vehicle type can be almost correctly determined. FIG. 10 shows the relationship among the vehicle region T, the model region M, and the intersection area C in this case. However, if the vehicle in the lane to be determined passes through the detection area D while there is a vehicle passing in the lane parallel to this lane, the lane to be determined and the lane parallel to this lane are respectively set. The passing vehicles are extracted as one vehicle region T. For this reason,

(Equation 4) The vehicle type determination method in such a case will be described with reference to the flowchart in FIG.

Steps (1) to (11) correspond to FIG.
(1) to (11).

(1) An image of the detection area D is captured by the camera 1 (S21). The captured image is transmitted to the electronic device 2 by wire or wirelessly, and stored in the memory 11. When storing in the memory 11, the imaging time (storage time) when the image was captured (stored)
Is stored in a format that also includes The camera 1 always captures the detection area D.

(2) Noise processing is performed on the captured image, and a background difference method (or an inter-frame difference method) is used.
Extract an area consisting of a plurality of pixels whose gray value has changed (S
twenty two). Then, the size of the extracted region is obtained. Most of the moving objects on the road are vehicles such as automobiles, and it is unlikely that humans or other objects exist. Therefore, the extracted area is regarded as the vehicle area T. The process of extracting the vehicle region T from the image is performed by the CPU 10, and the vehicle region T is stored in the memory 11 together with the size of the vehicle region T.

Note that, along with the extraction of the vehicle area T, the lane width on the image (see FIG.
5 (W1, W2)) in pixels (pixels), and the vehicle area T
And stored in the memory 11.

(3) The CPU 10 checks whether or not the vehicle area T is included in a specific area in the detection area D (S23). The specific region is a region indicated by diagonal lines in the perspective view of the detection region D indicating the specific region (see FIG. 6), and is a rectangular region having a width of approximately 2 meters in the traveling direction. If the front end (in the traveling direction) of the vehicle area T falls within this specific area, the rear end (in the traveling direction) enters the detection area D, and almost one vehicle is imaged. When the vehicle region T does not fall within the specific region, the process proceeds to S21 because there is no vehicle in the detection region D or the vehicle cannot be extracted. If the vehicle area T is within the specific area, the process proceeds to the next step.

(4) When a vehicle is present in the detection area D, a vehicle model of a large vehicle is read from the memory 11 among vehicle models stored for each vehicle type, and the read vehicle model and image An image is created by converting the width of the upper lane, the traveling direction of the vehicle, and the vehicle region T into the appearance in the vehicle region T of the vehicle model of the large vehicle. The area of the model area M (the number of pixels) is determined using the converted image of the vehicle model as the model area M (S24). Vehicle model conversion,
The area of the model area M is calculated and obtained by the CPU 10, and the model area M and the area are stored in the memory 11.

(5) The model area M is used as a mask pattern, the model area M is moved within the vehicle area T, and the maximum value of the intersection area C where the model area M and the vehicle area T intersect is obtained (S2).
Five). The movement of the model area M and the calculation of the intersection area C are performed by the CPU 10, and the intersection area is stored in the memory 11 in association with the vehicle model.
Store the maximum value of C.

(6) The CPU 10 checks whether or not the intersection area C has been obtained for all vehicle models (S26). Initially, only the intersection area C for a large vehicle is determined, so the process proceeds to the next step. If the intersection areas C for all vehicle models have been obtained, the process proceeds to (9).

(7) If the intersection area C has not been obtained for all vehicle models, the other vehicle models
For example, a vehicle model of an ordinary vehicle is read, and the model area M of the vehicle model of the ordinary vehicle and the area of the model area M are obtained (S27). The determined model area M and the area of the model area M are stored in the memory 11 in association with the vehicle model.

(8) The maximum value of the intersection area C between the model area M and the vehicle area T obtained in (7) is obtained (S28). Determined intersection area C
Is stored in the memory 11 in association with the vehicle model. Proceed to (S26).

(9) If the vehicle area T is an arbitrary model area n (however,
A criterion of the accuracy (n) that is (n = 1, 2, 3, 4) is obtained and stored in the memory 11 (S29). This accuracy (n) is defined as C / M.

The accuracy (n) is that if the selected vehicle model is correct for the vehicle region T, the size of the intersection area C approaches the size of the model region M,

(Equation 5) Similarly, if the selected model area M is smaller than the vehicle area T,

(Equation 6) On the other hand, if the selected model area M is larger than the vehicle area T,

(Equation 7) (10) T / M is obtained for every vehicle model (S30). Asked
T / M is stored in the memory 11 in association with the vehicle model, and this calculation is performed by the CPU 10.

(11) A new accuracy (n) is obtained for each vehicle model based on a predetermined relationship and stored (S31). The predetermined relationship is accuracy if T / M obtained in (S30) is T / M <1
(n) = Accuracy (n) × M / T is used, and if T / M ≧ 1, it means using the equation of accuracy (n) = Accuracy (n) × T / M.

The calculation of the new accuracy (n) is performed by the CPU 10, and the new accuracy (n) is stored in the memory 11.

(12) Intersection area determined for each vehicle model
A ratio (C / T) between C and the vehicle area T is determined (S32). The C / T is calculated by the CPU 10, and the obtained C / T is stored in the memory 11 in association with the vehicle model.

(13) Compare C / T with threshold value 1 (TH1) (S33).
The threshold value 1 is set to, for example, 0.4, but this numerical value can be appropriately set by the user. When C / T is smaller than the threshold value 1, the accuracy (n) of the vehicle model is multiplied by a predetermined amount (0.4) and reduced. After the new accuracy (n) is obtained by reducing the accuracy (n) by a predetermined amount, and when the C / T is larger than the threshold 1, the process proceeds to the next step (14).

By this step, the vehicle to be extracted
When a vehicle running in parallel with the vehicle to be extracted is extracted as one vehicle region T, it is possible to prevent the vehicle to be extracted from being determined to be a larger vehicle type, and to perform almost correct vehicle type determination. For this reason, it is possible to prevent erroneous discrimination such as discriminating a small car from a normal car and a large car, for example.

It should be noted that the numerical value of the predetermined amount to be multiplied by the accuracy (n) can be appropriately set by the user.

(14) Intersection area determined for each vehicle model
The ratio (C / M) between C and the model area M is determined (S34). C / M is CPU1
The C / M calculated and calculated based on 0 is stored in the memory 11 in correspondence with the vehicle model.

(15) The C / M is compared with the threshold value 2 (TH2) (S35).
The threshold value 2 is set to, for example, 0.9, but this numerical value can be appropriately set by the user. When C / M is larger than the threshold value 2, the accuracy (n) of the vehicle model is multiplied by a predetermined amount (1.5) to increase. After the accuracy (n) is increased by a predetermined amount to obtain a new accuracy (n), and when the C / M is smaller than the threshold value 2, the process proceeds to the next step (16).

It should be noted that the numerical value of the predetermined amount by which the accuracy (n) is multiplied can be appropriately set by the user.

(16) Accuracy (n) obtained as described above
Is used to determine the vehicle type (S36). Accuracy (n)
It is assumed that the vehicle model having the maximum value of is a vehicle included in the vehicle region T captured by the camera 1. Memory 11
The operation of reading the accuracy (n) from, obtaining the maximum value, and determining the vehicle type is performed by the CPU. The determined vehicle type is stored in the memory 11 in association with the vehicle area T together with the imaging time.

Through the steps (1) to (16) as described above, the camera 1 determines which type of vehicle is in the captured image. This method is particularly effective when a parallel vehicle is extracted as the vehicle region T (when the frequency of extraction is high) and when it is desired to distinguish the large vehicle from a large vehicle.

Next, when the model area M is smaller than that of a large vehicle such as a motorcycle, if the vehicle area T is smaller in proportion to the size of the model area M, the vehicle type can be almost correctly determined. Vehicle area T, model area M, intersection area C at this time
Is shown in FIG. However, if the vehicle area T is smaller than the model area M,

(Equation 8) Steps (1) to (11) are the same as (1) to (11) in FIG. 3 described above.

(1) An image of the detection area D is captured by the camera 1 (S41). The captured image is transmitted to the electronic device 2 by wire or wirelessly, and stored in the memory 11. When storing in the memory 11, the imaging time (storage time) when the image was captured (stored)
Is stored in a format that also includes The camera 1 always captures the detection area D.

(2) Noise processing is performed on the captured image, and the background difference method (or the inter-frame difference method) is used.
Extract an area consisting of a plurality of pixels whose gray value has changed (S
42). Then, the size of the extracted region is obtained. Most of the moving objects on the road are vehicles such as automobiles, and it is unlikely that humans or other objects exist. Therefore, the extracted area is regarded as the vehicle area T. The process of extracting the vehicle region T from the image is performed by the CPU 10, and the vehicle region T is stored in the memory 11 together with the size of the vehicle region T.

In addition to the extraction of the vehicle area T, the lane width (see FIG. 5 (W
1, W2)) are obtained in units of pixels (pixels), and stored in the memory 11 in association with the vehicle region T.

(3) The CPU 10 checks whether or not the vehicle area T is included in a specific area in the detection area D (S43). The specific region is a region indicated by diagonal lines in the perspective view of the detection region D indicating the specific region (see FIG. 6), and is a rectangular region having a width of approximately 2 meters in the traveling direction. If the front end (in the traveling direction) of the vehicle area T falls within this specific area, the rear end (in the traveling direction) enters the detection area D, and almost one vehicle is imaged. When the vehicle region T does not fall within the specific region, the process proceeds to S1 because there is no vehicle in the detection region D or the vehicle cannot be extracted. If the vehicle area T is within the specific area, the process proceeds to the next step.

(4) When a vehicle exists in the detection area D, a vehicle model of a large vehicle is read out of the memory 11 among the vehicle models stored for each vehicle type, and the read vehicle model and lane are read. From the width of the vehicle, the traveling direction of the vehicle and the
An image converted to the appearance in the vehicle region T of the vehicle model of the large vehicle is created. Then, the area of the model area M (the number of pixels) is determined using the converted image of the vehicle model as the model area M (S44). The conversion of the vehicle model and the area of the model area M are calculated and obtained by the CPU 10, and the model area M and the area are stored in the memory 11.

(5) The model area M is used as a mask pattern, the model area M is moved within the vehicle area T, and the maximum value of the intersection area C where the model area M and the vehicle area T intersect is obtained (S4).
Five). The movement of the model area M and the calculation of the intersection area C are performed by the CPU 10, and the intersection area is stored in the memory 11 in association with the vehicle model.
Store the maximum value of C.

(6) The CPU 10 checks whether or not the intersection area C has been obtained for all vehicle models (S46). Initially, only the intersection area C for a large vehicle is determined, so the process proceeds to the next step. If the intersection areas C for all the vehicle models have been obtained, the process proceeds to (9).

(7) If the intersection area C has not been obtained for all the vehicle models, the other vehicle models,
For example, a vehicle model of an ordinary vehicle is read, and the model area M and the area of the model area M of the read ordinary vehicle model are obtained (S47). The determined model area M and the area of the model area M are stored in the memory 11 in association with the vehicle model.

(8) The maximum value of the intersection area C between the model area M and the vehicle area T obtained in (7) is obtained (S48). Determined intersection area C
Is stored in the memory 11 in association with the vehicle model. Proceed to (S46).

(9) The vehicle area T is an arbitrary model area n (however,
A criterion for the accuracy (n) that is (n = 1, 2, 3, 4) is obtained and stored in the memory 11 (S49). This accuracy (n) is defined as C / M.

When the selected vehicle model is correct for the vehicle region T, the accuracy (n) is such that the size of the intersection area C approaches the size of the model region M,

(Equation 9) Similarly, if the selected model area M is smaller than the vehicle area T,

(Equation 10) On the other hand, if the selected model area M is larger than the vehicle area T,

[Equation 11] (10) T / M is obtained for each vehicle model (S50). Asked
T / M is stored in the memory 11 in association with the vehicle model, and this calculation is performed by the CPU 10.

(11) A new accuracy (n) is obtained for each vehicle model based on a predetermined relationship and stored (S51). The predetermined relationship is accuracy if T / M obtained in (S50) is T / M <1
(n) = Accuracy (n) × M / T is used, and if T / M ≧ 1, it means using the equation of accuracy (n) = accuracy (n) × T / M.

The calculation of the new accuracy (n) is performed by the CPU 10, and the new accuracy (n) is stored in the memory 11.

(12) For each vehicle model obtained in (10)
The size of T / M is checked (S52). When the magnitude of T / M is 1 or less, the accuracy (n) of the vehicle model is increased by multiplying by a predetermined amount (1.5) and stored in the memory 11. If the magnitude of T / M is greater than 1, the accuracy (n) is left as it is.

The numerical value which is a predetermined amount to be multiplied by the accuracy (n) can be appropriately set by the user.

(13) Accuracy (n) obtained as described above
Is used to determine the vehicle type (S53). Accuracy (n)
It is assumed that the vehicle model having the maximum value of is the vehicle included in the vehicle region T captured by the camera 1. Memory 11
The operation of reading the accuracy (n) from, obtaining the maximum value, and determining the vehicle type is performed by the CPU. The determined vehicle type is stored in the memory 11 in association with the vehicle area T together with the imaging time.

According to the steps (1) to (13) described above, the camera 1 determines which type of vehicle is in the captured image. This method is particularly effective when discriminating a small car or a motorcycle.

Further, as shown in the flow chart of FIG. 15, it may have a step that simultaneously includes the steps of FIG. 12 (when the model area is large) and 14 (when the model area is small).

In the first embodiment as described above, in order to extract the vehicle region T, the vehicle model is not a real image but a 3
By using the dimensional shape model, real-time processing is enabled, restrictions on the installation conditions of the camera 1 are reduced, and even when a plurality of vehicles are imaged in an overlapping manner, the vehicle type can be almost correctly determined.

Further, since the reliability of the determination can be enhanced, it can be used for wide area monitoring of expressways and general roads.

Next, the configuration of the second embodiment of the present invention will be described with reference to the block diagram of the second embodiment of FIG.

In the following embodiments, the same components are denoted by the same reference numerals, and duplicate description will be omitted.

The feature of the second embodiment is that a model area storage device 20 is provided in the electronic device 2.

In the electronic device 2, a model area storage device 20 is arranged so as to be connected to the model area projection processing device 6 and the discriminating device 7.

The model area storage device 20 stores a model area M for each vehicle model in advance. Model area M
Is the same as in the first embodiment.

In the first embodiment, since the positional relationship between the camera and the vehicle changes, it is necessary to convert the size and appearance of the vehicle model corresponding to the positional relationship. That is,
If the distance between the camera and the vehicle is small, the image of the vehicle is large, and if it is large, the image of the vehicle is small. Therefore, the model area M cannot be calculated and stored in advance for each vehicle model. If the model region M calculated in advance is used for the discrimination, the discrimination cannot be performed or the vehicle type cannot be discriminated correctly.

On the other hand, in the second embodiment, when the camera 1 is located sufficiently far (15 to 20 meters or more) from the detection area D, the vehicle existing in the detection area D ,
It can be approximated to one model area M regardless of the position in the detection area D. Therefore, the model area storage device 2
At 0, the model area M for each vehicle type can be calculated and stored in advance.

The vehicle type discriminating method is the same as in the first embodiment.

In the second embodiment as described above,
Depending on the installation position of camera 1, model area for each vehicle model
M can be calculated beforehand, and the calculation cost can be reduced. The second embodiment is more advantageous on a highway and the like than on a general road because it can cope with real-time processing and has a low calculation cost.

The present invention is not limited to the above-described embodiments, and it goes without saying that various modifications can be made without departing from the spirit of the present invention. For example, when a plurality of vehicles are captured in an image captured by the camera, the plurality of vehicles may be extracted as one vehicle region. If the vehicle region can be divided into a plurality of vehicle regions, the vehicle region may be divided and a plurality of vehicle regions may be extracted.

As a method of extracting the vehicle area,
It is also possible to use optical flow or color information, and the image captured by the camera is captured using visible light or infrared light.

Further, the number of vehicle types to be determined may be any number. For example, when five types (large-sized car, medium-sized car, ordinary car, light car (small-sized car), two-wheeled vehicle) are used, a medium-sized car By model

(Equation 12) When, the accuracy (n) of the medium-sized vehicle may be reduced, and the accuracy (n) of the large-sized vehicle may be further reduced (compared to the medium-sized vehicle). It is also possible to reduce only the accuracy (n) of a large vehicle.

Further, in a vehicle model of a medium-sized car,

(Equation 13) , The accuracy (n) of the medium-sized vehicle may be increased, and the accuracy (n) of the large-sized vehicle may be further increased (compared to the medium-sized vehicle). It is also possible to increase only the accuracy (n) of a large vehicle.

Further, in a vehicle model of a large car,

[Equation 14] , Only the accuracy (n) of the large vehicle needs to be increased.

Further, the processing in the embodiment of the present invention can be realized by a computer-executable program, and the program can be realized as a computer-readable storage medium.

The storage medium of the present invention includes:
Magnetic disk, floppy (registered trademark) disk, hard disk, optical disk (CD-ROM, CD-R, DVD, etc.),
As long as the program can be stored and a computer-readable storage medium such as a magneto-optical disk (MO or the like), a semiconductor memory, or the like can be used, any storage format may be used.

An OS (operation system) running on the computer, MW (middleware) such as database management software, network, etc., according to the instructions of the program installed in the computer from the storage medium, etc. A part of each process for realizing it may be executed.

Further, the storage medium in the present invention is not limited to a medium independent of a computer, but also includes a storage medium in which a program transmitted through a LAN, the Internet, or the like is downloaded and stored or temporarily stored.

Further, the number of storage media is not limited to one, and a case where the processing in the present embodiment is executed from a plurality of media is also included in the storage medium of the present invention, and the configuration of the medium may be any configuration. Good.

The computer according to the present invention executes each processing in the present embodiment based on a program stored in a storage medium.
Any configuration, such as a single device or a system in which a plurality of devices are connected to a network, may be used.

The computer in the present invention is
It is not limited to a personal computer, but also includes a processing device, a microcomputer, and the like included in an information processing device, and generically refers to a device and a device capable of realizing the functions of the present invention by a program.

Further, according to the present invention, the object to be determined is an automobile (2
Wheel, four wheels), but it is also possible to apply to other moving objects such as airplanes, trains, ships, bicycles and humans.

[0109]

As described above, according to the present invention, real-time processing is possible, and a desired vehicle type can be determined with less restrictions on camera installation conditions.

[Brief description of the drawings]

FIG. 1 is a perspective view of a first embodiment of a vehicle type identification device according to the present invention when installed on a road.

FIG. 2 is a block diagram of a first embodiment of a vehicle type identification device according to the present invention.

FIG. 3 is a block diagram of an electronic device of the vehicle type identification device of the present invention.

FIG. 4 is a flowchart of the first embodiment of the vehicle type identification device of the present invention.

FIG. 5 is an explanatory diagram of the operation of the vehicle type discriminating method of the present invention.

FIG. 6 is an explanatory diagram of the operation of the vehicle type discrimination method of the present invention.

FIG. 7 is an explanatory diagram of the operation of the vehicle type discriminating method of the present invention.

FIG. 8 is an explanatory diagram of the operation of the vehicle type discrimination method of the present invention.

FIG. 9 is an explanatory diagram of the operation of the vehicle type determination method of the present invention.

FIG. 10 is an explanatory diagram of the operation of the vehicle type discrimination method of the present invention.

FIG. 11 is an explanatory diagram of the operation of the vehicle type discriminating method of the present invention.

FIG. 12 is a flowchart of a vehicle type determination method according to the present invention.

FIG. 13 is an explanatory diagram of the operation of the vehicle type discriminating method of the present invention.

FIG. 14 is a flowchart of a vehicle type determination method according to the present invention.

FIG. 15 is a flowchart of a vehicle type determination method according to the present invention.

FIG. 16 is a block diagram of a second embodiment of the vehicle type identification device of the present invention.

[Explanation of symbols]

 1 Camera 2 Electronic equipment 3 Post 4 Image signal processing device 5 Model storage device 6 Model projection processing device 7 Classification device 20 Model region storage device

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G08G 1/04 G08G 1/04 DF Term (Reference) 5B057 AA16 BA02 CA08 CA12 CA16 DA12 DB02 DB09 DC33 5H180 AA01 CC04 EE07 EE10 5L096 AA06 BA04 CA02 FA02 HA03 HA09 JA11

Claims (6)

[Claims] 1. An image pickup means for picking up an image in which an area where a moving object passes is set as an imaging range, and a moving object area extracting means for extracting an area including the moving object from the picked-up image as a moving object area. A storage unit that stores a three-dimensional shape of a plurality of types of moving objects as a moving object model for each vehicle type of the moving object; and a movement that converts the moving object model into a view of the imaging direction of the moving object on the image. A moving body model image converting means for converting the moving body model image into a body model image, and a vehicle type for comparing the degree of overlap when the moving body model image is overlaid on the moving body on the image, and discriminating the vehicle type of the moving body from the comparison result. A vehicle type discriminating device comprising a discriminating means. 2. The moving body model is data representing a characteristic of a three-dimensional shape of the moving body using a plurality of points,
2. The vehicle type discrimination device according to claim 1, wherein the moving object model image is data that connects the plurality of points and expresses the two-dimensionally. 3. The vehicle type discriminating means superimposes the moving body region and the moving body model image to obtain an area of a region where the moving body region intersects with the moving body model image. 2. The vehicle type determination device according to claim 1, wherein the vehicle type of the mobile object is determined from a size of the mobile object region and a size of the mobile object model image. 4. A method of capturing an image in which an area where a moving object passes is defined as an imaging range; a step of extracting a region including the moving object from the captured image as a moving object area; Storing the three-dimensional shape of the body as a moving body model for each vehicle type of the moving body; and converting the moving body model into a moving body model image converted into a view of the moving body in the imaging direction on the image. A moving object model image converting step, and a step of comparing the degree of overlap when the moving object model image is superimposed on the image, and discriminating a vehicle type of the moving object from the comparison result. Method. 5. A storage medium in which a program for determining the type of a vehicle is stored in a computer-readable manner, wherein an image having an area where a moving object passes is taken as an imaging range. A region including the moving object is extracted from the inside as a moving object region, and a three-dimensional shape of a plurality of types of moving objects is stored as a moving object model for each vehicle type of the moving object. Converted to a moving body model image converted to the appearance of the imaging direction of the moving body,
A computer-readable storage medium storing a program for comparing a degree of overlap when the moving object model image is superimposed on the image and determining a vehicle type of the moving object from the comparison result. 6. A function for causing a computer to capture an image in which an area where a moving object passes is set as an imaging range; a function for extracting a region including the moving object from the imaged image as a moving object area; A function of storing a three-dimensional shape of the moving object as a moving object model for each vehicle type of the moving object; and a moving object model obtained by converting the moving object model into a view of the imaging direction of the moving object on the image. A computer program having a function of converting an image into an image and a function of comparing the degree of overlap when the moving object model image is superimposed on the image and determining the vehicle type of the moving object based on the comparison result.