JP2006163726A - Vehicle perimeter recognition system and image processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle perimeter recognition system for recognizing vehicle perimeter such as obstacles and passengers by photographing the outer field of the vehicle and an image processing system constructing the vehicle perimeter recognition system. <P>SOLUTION: An ECU2 carries out two dimensional Fourier transformation of one frame photographed image photographed by a video camera 1a, and transmits two dimensional Fourier transformed image to an ECU3, ECU4 and ECU5. The ECU2 assigns standard patterns H1, H2, ...Hx to the ECU2, ECU3, ECU4, and ECU5, respectively. Then, the ECU2, ECU3, ECU4 and ECU5 carry out sum of products (convolution) of two dimensional Fourier transformed image and conjugate complex numbers of Fourier transformed standard patterns, and two dimensional inverse Fourier transformation in parallel, thereby distributing computation throughput on the respective ECU2, ECU3, ECU4 and ECU5. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a vehicle periphery recognition system that images the outside of a vehicle and recognizes the periphery of the vehicle such as an obstacle or a pedestrian, and an image processing apparatus that constitutes the vehicle periphery recognition system.

  A far-infrared video camera equipped with a bolometer or pyroelectric image sensor is mounted on a vehicle such as an automobile, and images taken of obstacles and pedestrians in front of the vehicle with the video camera during night driving. A monitoring system has been proposed in which an image processing ECU is used to recognize obstacles, pedestrians, etc., and alert the driver.

Such a system uses a fact that when a pedestrian is imaged with a far-infrared video camera, the pedestrian, which is a heat source, has a high luminance with respect to other background objects, and based on the eigenspace method in advance. A reference pattern representing the characteristics of a pedestrian is stored, a correlation calculation is performed between the captured image captured by the far-infrared video camera and the reference pattern, and the captured image is based on the result of the correlation calculation. A pedestrian or the like is recognized by judging whether or not it is similar to a standard pattern (see Patent Document 1).
JP 2001-201575 A

 However, in the example of Patent Document 1, the correlation calculation of the eigenvector between the captured image and the standard pattern stored in advance requires a very large calculation processing capability. In addition, in order to recognize a pedestrian with high accuracy, it is necessary to store a large number of standard patterns and repeat the correlation operation as many times as the number of standard patterns in order to determine which standard pattern the captured image is similar to. is there. For this reason, in order to recognize a pedestrian in real time, a high-speed DSP (Digital Signal Processor) or a large-scale LSI is required, and there is a problem that the cost of the ECU for image processing increases.

  The present invention has been made in view of such circumstances, and includes a plurality of image processing devices, transmits captured images obtained by imaging with the imaging device to other image processing devices, and each image processing device. Using different standard patterns, the periphery of the vehicle is recognized based on the result of determination by the determination means for determining whether a captured image received by another image processing apparatus is similar to the standard pattern. Accordingly, an object of the present invention is to provide a vehicle periphery recognition system capable of reducing the processing capacity required for each image processing device and reducing the cost as compared with the prior art, and an image processing device constituting the vehicle periphery recognition system. To do.

  Another object of the present invention is that the image processing apparatus that has transmitted the captured image determines whether or not the captured image is similar to the standard pattern. It is an object of the present invention to provide a vehicle periphery recognition system that can reduce the necessary processing capacity and reduce the cost as compared with the conventional art.

  Another object of the present invention is that it is necessary for recognition processing around the vehicle by performing parallel processing to determine whether or not a captured image obtained by imaging is similar to the standard pattern. An object of the present invention is to provide a vehicle periphery recognition system that can reduce the time required.

  A vehicle periphery recognition system according to a first aspect of the present invention captures an external environment of a vehicle with an imaging device, and processes the captured image obtained by imaging and a plurality of stored standard patterns with an image processing device to recognize the vehicle periphery. In the vehicle periphery recognition system, the image processing apparatus includes a plurality of the image processing apparatuses, the transmission / reception means for transmitting and receiving captured images obtained by capturing images to and from other image processing apparatuses, and each image processing apparatus. A judgment unit that judges whether a captured image received by the transmission / reception unit is similar to the standard pattern using a different standard pattern, and the image processing apparatus captures a captured image obtained by imaging Is transmitted to another image processing apparatus, and the periphery of the vehicle is recognized based on the result of determination by the determination means of the other image processing apparatus.

  In the vehicle periphery recognition system according to a second aspect, in the first aspect, the determination means of the image processing apparatus that has transmitted the captured image determines whether the captured image obtained by capturing is similar to the standard pattern. The image processing apparatus is characterized in that the periphery of the vehicle is recognized on the basis of the determination result of the determination means and the determination means of another image processing apparatus.

  A vehicle periphery recognition system according to a third aspect of the present invention is the first or second aspect of the present invention, wherein the determination whether or not the captured image obtained by imaging is similar to the standard pattern is processed in parallel. It is characterized by that.

  An image processing apparatus according to a fourth aspect of the present invention is an image processing apparatus for recognizing the periphery of a vehicle by processing a captured image and a plurality of stored standard patterns, and transferring the captured image to another image processing apparatus. A transmission / reception unit that transmits and receives; and a determination unit that determines whether a captured image received by the transmission / reception unit is similar to the standard pattern by using a different standard pattern for each of the image processing apparatuses. The vehicle periphery is recognized based on the result determined by the means.

  In the first invention and the fourth invention, the image processing apparatus having a plurality of image processing apparatuses and acquiring the captured image obtained by imaging with the imaging apparatus transmits the captured image to another image processing apparatus. . Using different standard patterns for each image processing apparatus, other image processing apparatuses determine whether the received captured image is similar to the standard pattern and transmit the determination result to the captured image Output to. The image processing apparatus recognizes the periphery of the vehicle based on the output determination result.

  In the second invention, the image processing apparatus that has transmitted the captured image determines whether or not the captured image obtained by imaging is similar to the standard pattern. The image processing device recognizes the periphery of the vehicle based on the result determined by the determining means and the result determined by another image processing device.

  In the third aspect of the invention, the determination whether or not the captured image obtained by imaging is similar to the standard pattern is processed in parallel.

  In the first invention and the fourth invention, it is determined whether or not the captured image obtained by imaging in each of the plurality of image processing apparatuses is similar to a standard pattern that differs for each image processing apparatus. By recognizing the vehicle periphery based on the result, the recognition processing of the vehicle periphery can be distributed by a plurality of image processing devices, and the vehicle periphery recognition is not required, unlike the conventional high-speed or large-scale image processing device. The cost of the system can be reduced.

  In the second invention, in addition to other image processing devices, the image processing device that has transmitted the captured image can further disperse the processing by making a similarity determination between the captured image and the standard pattern. The processing capability required for each image processing apparatus can be reduced to lower the cost.

  According to the third aspect of the present invention, it is possible to reduce the time required for the recognition processing around the vehicle in parallel by processing the similarity determination between the captured image obtained by imaging and the standard pattern in parallel. It is possible to recognize the vicinity of the vehicle.

  Hereinafter, the present invention will be described with reference to the drawings illustrating embodiments. FIG. 1 is a schematic diagram showing an outline of a vehicle periphery recognition system according to the present invention. In the figure, reference numeral 1a denotes a far-infrared video camera that is mounted on a vehicle and recognizes a pedestrian at night, a person on a bicycle, and the like. The video camera 1a is disposed on the front grille, and is connected to the pedestrian recognition ECU 2 via an in-vehicle LAN communication line 7 compliant with IEEE1394.

  The pedestrian recognition ECU 2 is connected to far-infrared video cameras 1b, 1c, and 1d via a communication line 7. The video camera 1b is installed in a bumper at the rear of the vehicle and recognizes an obstacle behind the vehicle. The video camera 1c is installed under the left door mirror of the vehicle and recognizes an approaching vehicle, an obstacle, and the like on the left side of the vehicle. The video camera 1d is installed below the right door mirror of the vehicle and recognizes an approaching vehicle, an obstacle, and the like on the right side of the vehicle.

  Further, the pedestrian recognition ECU 2 includes an obstacle recognition ECU 3, an approaching vehicle recognition ECU 4, a white line recognition ECU 5, an alarm unit and an operation unit that issue a warning by voice or sound effect via a communication line 7. Is connected.

  By connecting each ECU via the communication line 7, for example, the pedestrian recognition ECU 2 performs a pedestrian recognition process based on a captured image captured by the video camera 1 a, and an obstacle recognition ECU 3. Performs the obstacle recognition process based on the captured images captured by the video cameras 1b, 1c, and 1d, and the approaching vehicle recognition ECU 4 performs the approaching vehicle based on the captured images captured by the video cameras 1c and 1d. The white line recognition ECU 5 performs white line recognition processing based on the captured images captured by the video cameras 1c and 1d.

  FIG. 2 is a block diagram showing the configuration of the video camera 1a. Since the video cameras 1b, 1c, and 1d have the same configuration, description thereof is omitted. In the figure, reference numeral 11 denotes an image capturing unit, in which a far-infrared image sensor that converts an optical signal into an electrical signal is arranged, and the far-infrared image sensor converts input far-infrared light according to the intensity of far-infrared light. Is converted to a luminance signal and output to the signal processing unit 12.

  The signal processing unit 12 performs processing for removing various distortions generated in the optical system, processing for removing low-frequency noise, correction processing for correcting gamma characteristics, and the like on the luminance signal input from the image capturing unit 11. The processed luminance signal is temporarily stored in the image memory 13 as image data.

  The interface unit 14 communicates with the ECUs 2, 3, 4, and 5 through the communication line 7, and the captured images stored in the image memory 13 are transmitted to the ECUs 2, 3, 3 according to instructions transmitted from the ECUs 2, 3, 4, and 5. 4 and 5, conversion of the transfer rate according to the resolution of the image captured by the video camera 1 a, creation of packet data from the captured image, and the like. Further, the interface unit 14 outputs the captured image stored in the image memory 13 to the ECUs 2, 3, 4, and 5 together with the identification number of the video camera 1a under the control of the control unit 15.

  The control unit 15 controls processing of the image capturing unit 11, the signal processing unit 12, and the interface unit 14. For example, the command from the ECU 2 is interpreted, and the captured image captured by the image capturing unit 11 is stored in the image memory 13. The stored captured image is read out and output to the ECU 2 via the interface unit 14.

  FIG. 3 is a block diagram showing the configuration of the ECU 2. In the figure, reference numeral 21 denotes an interface unit that transmits commands to the video cameras 1a, 1b, 1c, and 1d, and receives captured images from the video cameras 1a, 1b, 1c, and 1d. The interface unit 21 communicates with the other ECUs 3, 4, 5 and transmits / receives commands or data to / from the other ECUs 3, 4, 5.

  The input unit 25 receives a sensor signal from a sensor (for example, a speed sensor) mounted on the vehicle, a signal from an operation unit (for example, a pedestrian recognition process start switch) provided in the display device 6, and the like. Input unit 25 outputs the received signal to control unit 27.

  The control unit 27 stores the captured images input from the video cameras 1a, 1b, 1c, or 1d via the interface unit 21 in the image memory 22 for each frame unit. Further, the control unit 27 outputs the captured image stored in the image memory 22 to the image processing LSI 26 in units of frames.

  Based on the signal input to the input unit 25, the control unit 27 loads one of the recognition processing units stored in the storage unit 23, which will be described later, into the image processing LSI 26, so that the image processing LSI 26 performs predetermined processing. Control to do. For example, when the driver starts the pedestrian recognition process by operating the operation unit provided in the display device 6, the control unit 27 is stored in the storage unit 23 based on the signal input from the input unit 25. The pedestrian recognition processing unit 231 loaded on the image processing LSI 26 is loaded.

  The control unit 27 assigns the standard patterns H1, H2,... Hx stored in the storage unit 23 to the ECUs 2, 3, 4, 5 almost evenly. Further, the control unit 27 compares the degree of correlation calculated by the image processing LSI 26 with the threshold value TH stored in the storage unit 23 to extract pedestrians.

  The storage unit 23 includes a pedestrian recognition processing unit 231 that identifies a process for recognizing a pedestrian, an obstacle recognition processing unit 232 that identifies a process for recognizing an obstacle, and a process for recognizing an approaching vehicle. An approaching vehicle recognition processing unit 233 to be specified and a white line recognition processing unit 234 to specify a process for recognizing a white line on the road surface are stored.

  The memory | storage part 23 has memorize | stored the standard pattern part 235 which has the standard patterns H1, H2, ... Hx which show the luminance distribution of a pedestrian. Standard patterns H1, H2,... Hx are a plurality of different patterns depending on the pedestrian's body shape, the distinction between children and adults, clothing, the pedestrian's posture, position, and the like.

  The standard pattern unit 235 includes a standard pattern for recognizing an obstacle, a standard pattern for recognizing an approaching vehicle, and a standard pattern for recognizing a white line.

  The image processing LSI 26 is configured by an FPGA (Field Programmable Gate Array) incorporating a logic block, a memory, an internal register, and the like. The FPGA loads a logical bit string (architecture bit) specified by each recognition processing unit stored in the storage unit 23 into each built-in logical block or memory, thereby enabling a pedestrian recognition process using a matched filter. I do. That is, a product-sum operation is performed on a captured image that has been subjected to two-dimensional Fourier transform and a conjugate complex number of Fourier transform of a standard pattern, and the degree of correlation (correlation power) is calculated by performing inverse two-dimensional Fourier transform on the product-sum result. The maximum value of the correlation degree is output to the control unit 27. Further, the FPGA loads the logical bit string specified by the obstacle recognition processing unit 232, the approaching vehicle recognition processing unit 233, or the white line recognition processing unit 234, so that the obstacle recognition process, the approaching vehicle recognition process, or the white line recognition is performed. Processing can also be performed.

  The image processing LSI 26 reads out the captured image stored in the image memory 22 in units of frames, and outputs the captured image as a result of performing the pedestrian recognition process, the obstacle recognition process, the approaching vehicle recognition process, or the white line recognition process. Output to. The image processing LSI 26 repeats the same processing for each captured image stored in the image memory 22 in units of frames.

  The output unit 24 transmits the processed captured image output from the image processing LSI 26 to the display device 6.

  The display device 6 displays the situation around the vehicle on the display based on the output captured image and, when a pedestrian is recognized, attaches a frame around the recognized pedestrian to the driver. Highlight to encourage the presence of pedestrians. In addition, recognized approaching vehicles, obstacles, white lines, etc. are displayed.

  The alarm unit provided in the display device 6 emits a sound or a warning sound when a pedestrian, an approaching vehicle, or an obstacle is recognized, and alerts the driver.

  Since the ECUs 3, 4, and 5 have the same configuration, the description thereof is omitted. The ECU 2 is loaded with a logical bit string specified by the pedestrian recognition processing unit 231 in the image processing LSI 26 by default, and performs image processing for pedestrian recognition. By default, the ECU 3 is loaded with a logical bit string specified by the obstacle recognition processing unit 232 in the image processing LSI, and performs image processing for obstacle recognition. The ECU 4 is loaded with a logical bit string specified by the approaching vehicle recognition processing unit 233 in the image processing LSI by default, and performs image processing for approaching vehicle recognition. The ECU 5 loads the logical bit string specified by the white line recognition processing unit 234 to the image processing LSI by default, and performs image processing for white line recognition.

  FIG. 4 is an explanatory diagram showing a configuration of instructions used between ECUs. The command is created by the control unit 27, and is composed of a transmission source address that identifies the transmission source ECU, a transmission destination address that identifies the transmission destination ECU, and a data portion to be transmitted. The data part is composed of a command part for the destination ECU and command data used by the command part.

  The command unit requests communication start with the transmission destination ECU, and confirms whether the transmission destination ECU can respond to the request, “communication start” command, standard patterns H1 and H2 included in the standard pattern unit 235, ... "Standard pattern setting" command for assigning Hx to each ECU, "pedestrian recognition start" command for requesting other ECUs to start pedestrian recognition processing, and for ending pedestrian recognition processing of other ECUs It has a “pedestrian recognition end” command and the like.

  For example, when the instruction is “standard pattern setting”, the control unit 27 creates a standard pattern number assigned to each ECU as data. Thereby, the control unit 27 specifies which standard pattern is assigned to which ECU.

  FIG. 5 is an explanatory diagram showing an example of standard pattern assignment. As shown in the figure, the control unit 27 assigns the standard patterns H1, H2,... Hx used for pedestrian recognition almost evenly by the ECUs 2, 3, 4, and 5. Thereby, the ECU to which any one of the standard patterns H1, H2,... Hx is assigned performs pedestrian recognition processing using the assigned standard pattern.

  Next, pedestrian recognition processing will be described. When the driver operates the operation unit provided in the display device 6 to turn on the pedestrian recognition start switch, a pedestrian recognition start signal is input to the control unit 27 via the input unit 25 of the ECU 2. The The control unit 27 sends an initial command to the video camera 1a via the interface unit 21.

  Receiving the initial command, the video camera 1a starts imaging the outside of the vehicle from the image capturing unit 11. The signal processing unit 12 once records the captured image in the image memory 13 and then sends the captured image for one frame to the ECU 2 via the interface unit 14. The transmitted captured image is temporarily stored in the image memory. 22 is stored.

  When the control unit 27 receives a pedestrian recognition start signal, the control unit 27 transmits a “communication start” command to the other ECUs 3, 4, and 5. Thereby, it is confirmed whether other ECU3, 4, 5 is in the default recognition process. For example, when the driver has not started any obstacle recognition, approaching vehicle recognition, or white line recognition, the ECU 2 receives a confirmation signal that can perform a pedestrian recognition process from each of the ECUs 3, 4, and 5.

  The control unit 27 calculates the number of ECUs that can process the pedestrian recognition process in parallel based on the confirmation signals from the other ECUs 3, 4, 5, and based on the calculated number, the standard patterns H1, H2,. Decide the number of assignments to assign Hx almost evenly, determine the number of the standard pattern to be assigned to the ECUs 2, 3, 4, 5 based on the decided number, and set the determined standard pattern number together with the “standard pattern setting” command, It transmits to other ECU3,4,5.

  The control unit 27 outputs the captured image stored in the image memory 22 to the image processing LSI 26 in units of frames.

  The image processing LSI 26 performs a filtering process such as luminance correction on the received captured image for one frame, performs a two-dimensional Fourier transform on the captured image after the filtering process, and controls the captured image obtained by the two-dimensional Fourier transform. The product-sum operation (convolution operation) of the captured image obtained by two-dimensional Fourier transform and the conjugate complex number of the Fourier transform of the standard pattern assigned to the ECU 2 is performed for each standard pattern, and the product-sum results are reversed. The degree of correlation (correlation power) for each standard pattern is calculated by performing two-dimensional Fourier transform, and the maximum value M2 of the calculated degree of correlation is held in the internal register.

  The control unit 27 transmits the captured image obtained by the two-dimensional Fourier transform output from the image processing LSI 26 to the ECUs 3, 4, and 5 together with the “pedestrian recognition start” command.

  The ECU 3 that has received the “pedestrian recognition start” command standardizes the product-sum operation (convolution operation) of the received image after the two-dimensional Fourier transform and the conjugate complex number of the Fourier transform of the standard pattern assigned to the ECU 3. This is performed for each pattern, and the degree of correlation for each standard pattern is calculated by inverse two-dimensional Fourier transform of each product-sum result, and the calculated maximum value M3 of the degree of correlation is transmitted to the ECU 2.

  The ECU 4 that has received the “start pedestrian recognition” command standardizes the product-sum operation (convolution operation) of the received image after the two-dimensional Fourier transform and the conjugate complex number of the Fourier transform of the standard pattern assigned to the ECU 4. This is performed for each pattern, and the degree of correlation for each standard pattern is calculated by performing inverse two-dimensional Fourier transform on each product-sum result, and the maximum value M4 of the calculated degree of correlation is transmitted to the ECU 2.

  The ECU 5 that has received the “start pedestrian recognition” command standardizes the product-sum operation (convolution operation) of the received captured image after the two-dimensional Fourier transform and the conjugate complex number of the Fourier transform of the standard pattern assigned to the ECU 5. This is performed for each pattern, and the degree of correlation for each standard pattern is calculated by performing inverse two-dimensional Fourier transform on each product-sum result, and the calculated maximum value M5 of the degree of correlation is transmitted to the ECU 2. In addition, calculation of the correlation degree in ECU2,3,4,5 is processed in parallel. Thereby, ECU2 becomes a master and ECU3, 4, 5 becomes a slave, and the calculation calculation of the correlation degree for every standard pattern is processed in parallel.

  The control unit 27 of the ECU 2 extracts the maximum correlation degree value M2 held in the internal register of the image processing LSI 26, and also extracts the maximum correlation degree value M2 and the maximum correlation degree value M3 received from the ECUs 3, 4, and 5. , M4, M5 are compared with a predetermined threshold TH, and if any of the maximum correlation values M2, M3, M4, M5 is greater than the threshold TH, it is determined that there is a pedestrian and a pedestrian is extracted. To do.

  The image processing LSI 26 transmits the captured image from which the pedestrian has been extracted to the display device 6 via the output unit 24 based on the determination result of the control unit 27.

  The display device 6 displays the pedestrian's situation on the display based on the captured image output via the output unit 24, adds a frame around the extracted pedestrian, and presents the pedestrian to the driver. Highlight to prompt.

  When the driver operates the operation unit provided in the display device 6 to turn off the pedestrian recognition start switch, the control unit 27 issues a “pedestrian recognition end” command to the ECUs 3, 4, and 5. And the pedestrian recognition process in the ECUs 3, 4 and 5 is terminated.

  As described above, in the present invention, the standard patterns H1, H2,... Hx are assigned to the ECUs 2, 3, 4, and 5, and the ECUs 2, 3, 4, and 5 perform imaging after the two-dimensional Fourier transform. The calculation processing amount in each ECU 2, 3, 4, 5 is distributed by performing the product-sum operation (convolution operation) of the image and the conjugate complex number of the Fourier transform of the standard pattern and the inverse two-dimensional Fourier transform in parallel. And does not require a large-scale LSI. Further, since the arithmetic processing is performed in parallel, the processing time can be shortened and an expensive image processing device such as a high-speed DSP is not required. Further, since the ECUs 3, 4, and 5 perform dedicated processing by default, there is no need to newly add an ECU, and the cost of the entire system can be reduced by effectively using the existing ECU. .

  In the above-described embodiment, the pedestrian recognition process uses a matched filter. However, the present invention is not limited to this, and a configuration using another template matching method may be used. For example, a configuration in which a correlation value is calculated between the luminance values of the captured image and the reference pattern and a pedestrian is recognized based on the magnitude of the calculated correlation value may be used.

  In the above-described embodiment, the standard pattern is stored in the storage unit of the ECUs 2, 3, 4, and 5. However, the present invention is not limited to this, and is stored in one ECU (for example, only the ECU 2). The standard pattern may be stored in a storage device outside the ECU, and the assigned standard pattern may be read from the storage device and transmitted to each ECU. Thereby, the storage capacity for storing the standard pattern can be reduced. Further, a standard pattern for processing in each ECU may be distributed and stored in each ECU in advance.

  In the above-described embodiment, the standard patterns H1, H2,... Hx are stored. However, the complex complex number of the Fourier transform of the standard patterns H1, H2,.

  In the above-described embodiment, the ECUs 2, 3, 4, and 5 are connected via the communication line 7. However, the number of ECUs is not limited to this.

  In the above-described embodiment, the pedestrian recognition process is processed in parallel by each ECU. However, the present invention is not limited to this, and the process of recognizing the periphery of a vehicle such as an obstacle or an approaching vehicle is processed in parallel. It may be configured to. In this case, a standard pattern suitable for each process can be used.

  In the above-described embodiment, the connection between the video camera and the image processing apparatus is performed via the in-vehicle LAN communication line compliant with IEEE 1394. However, the present invention is not limited to this, and standards such as Gigabit Ethernet (registered trademark) are used. A compliant communication method may be used.

  In the above-described embodiment, the pedestrian recognition process is performed. However, the recognition target for recognizing the periphery of the vehicle is not limited to the pedestrian. The structure which recognizes the target object may be sufficient.

It is a schematic diagram which shows the outline | summary of the vehicle periphery recognition system which concerns on this invention. It is a block diagram which shows the structure of a video camera. It is a block diagram which shows the structure of ECU. It is explanatory drawing which shows the structure of the command used between ECU. It is explanatory drawing which shows the example of allocation of a standard pattern.

Explanation of symbols

1a, 1b, 1c, 1d Video camera 2, 3, 4, 5 ECU
6 Display Device 21 Interface Unit 22 Image Memory 23 Storage Unit 24 Output Unit 25 Input Unit 26 Image Processing LSI
27 Control Unit 231 Pedestrian Recognition Processing Unit 235 Standard Pattern

Claims (4)

In a vehicle periphery recognition system that captures an external environment of a vehicle with an imaging device, processes a captured image obtained by imaging and a plurality of stored standard patterns with an image processing device, and recognizes the vehicle periphery,
A plurality of the image processing devices;
The image processing apparatus includes:
A transmission / reception means for transmitting / receiving a captured image obtained by imaging to / from another image processing apparatus;
Using a different standard pattern for each of the image processing devices, and a determination unit that determines whether the captured image received by the transmission / reception unit is similar to the standard pattern,
The image processing apparatus includes:
Send the captured image obtained by imaging to another image processing device,
A vehicle periphery recognition system characterized in that the periphery of the vehicle is recognized based on a result determined by the determination means of the other image processing apparatus. The determination means of the image processing apparatus that has transmitted the captured image includes:
Whether or not a captured image obtained by imaging is similar to the standard pattern;
The image processing apparatus includes:
2. The vehicle periphery recognition system according to claim 1, wherein the vehicle periphery is recognized based on a result determined by the determination unit and a determination unit of another image processing apparatus.   The vehicle periphery recognition system according to claim 1 or 2, wherein a determination as to whether a captured image obtained by imaging is similar to the standard pattern is processed in parallel. In an image processing apparatus that recognizes the periphery of a vehicle by processing captured captured images and a plurality of stored standard patterns,
Transmission / reception means for transmitting / receiving the captured image to / from another image processing device;
Using a different standard pattern for each of the image processing devices, and a determination unit that determines whether the captured image received by the transmission / reception unit is similar to the standard pattern,
An image processing apparatus characterized by recognizing the periphery of a vehicle based on a result determined by the determining means.

Images