Patent Application
20210150229
This application is based on and claims priority under 35 USC 119 from Japanese Patent Application No. 2019-208883 filed on Nov. 19, 2019, the disclosure of which is incorporated by reference herein.
The present disclosure relates to an image processing system that processes an image, a processing device, a relay device, and a recording medium storing a program.
Japanese Patent Application Laid-open No. 2017-174244 discloses an information processing device intended to provide appropriate information to a vehicle that is a connected car, which is controlled based on recognition results obtained by external world recognition means.
Here, when image processing is performed on a captured image captured by a vehicle in the information processing apparatus disclosed in Japanese Patent Laid-Open No. 2017-174244 in order to identify a road obstacle, when image processing is performed in the information processing apparatus on all captured images transmitted from the vehicle side, the load at the side of the information processing apparatus is increased.
The present disclosure aims to provide an image processing system, a processing device, a relay device, and a recording medium storing a program, that can reduce the number of captured images to be processed by the processing-side device, and can reduce the processing load, when performing image processing for identifying an obstacle based on an image captured by a vehicle.
A first aspect is an image processing system that includes an image capture device provided in a vehicle, a transmission unit, a receiver unit, a calculation unit, and a processing unit. The transmission unit transmits, from the vehicle, a captured image captured by the image capture device as image information. The receiver unit receives image information transmitted from a plurality of the vehicles. The calculation unit calculates a degree of similarity between a captured image according. to newly received image information, and a reference image according to image information that has already been received. The processing unit performs image processing to identify an obstacle with respect to a captured image having a degree of similarity that is lower than a threshold value.
In the image processing system according to the first aspect, an image capture device and a transmission unit are provided in a vehicle, a receiver unit in a device outside the vehicle receives image information from a plurality of vehicles, a calculation unit calculates a degree of similarity between captured images, and a processing unit executes the image processing on a captured image having a degree of similarity degree that is lower than a threshold value.
Here, the degree of similarity of the captured image is calculated as a value with respect to a reference image that has already been received. In the image processing system, the configuration is such that a captured image having a degree of similarity that is equal to or higher than the threshold value is treated as an image that is similar to the reference image, and image processing is not executed. As a result, when performing image processing to identify an obstacle based on images captured at a vehicle, it is possible to reduce the number of captured images that are to be processed by the device that performs the image processing and the processing load can be reduced.
An image processing system according to a second aspect is the image processing system according to the first aspect, in which the receiver unit receives, together with the image information, first position information attached to the captured image and second position information on another vehicle, and further includes a determination unit and a notification unit. The determination unit determines whether or not the other vehicle having the second position information has approached the obstacle corresponding to the first position information. The notification unit notifies the other vehicle of the danger information in a case in which the determination unit determines that the other vehicle has approached the obstacle.
In the image processing system of the second aspect, the receiver unit receives the first position information of the captured image and the second position information of the other vehicle, and the determining unit determines whether the other vehicle has approached the Obstacle. In a case in which it is determined that the other vehicle has approached the obstacle, the notification unit notifies the other vehicle of the danger information. According to this image processing system, it is possible to alert another vehicle to an obstacle on the road based on the first position information of the obstacle.
An image processing system according to a third aspect is the image processing system according to the first or second aspect, including a control device, a relay device, and a processing device. The control device is installed at the vehicle having the transmission unit. The relay device includes the receiver unit, which receives image information transmitted from a plurality of the control devices, and the calculation unit, and transmits image information according to a degree of similarity that is lower than the threshold value. The processing device receives the image information transmitted from the relay device and includes the processing unit.
The image processing system according to the third aspect includes a control device installed at a vehicle, a processing device that executes image processing, and a relay device that performs relay between the control device and the processing device. According to this image processing system, since the relay device selects a captured image to be transmitted to the processing device based on the degree of similarity, it is possible to reduce the amount of communication in the communication line between the relay device and the processing device.
An image processing system according to a fourth aspect is the image processing system according to the third aspect, in which the relay device transmits image information according to a degree of similarity that is lower than the threshold value to the processing device and to discard image information according to a degree of similarity that is equal to or higher than the threshold value.
According to the image processing system of the fourth aspect, since an image that is deemed similar to the reference image is not stored in the relay device, resource consumption of the relay device can be suppressed.
A fifth aspect is a processing device that includes a receiver unit, a calculation unit, and a processing unit. The receiver unit receives, from a vehicle, image information according to a captured image captured by an image capture device provided at the vehicle. The calculation unit calculates a degree of similarity between a captured image according to newly received image information, and a reference image according to image information that has already been received. The processing unit performs image processing to identify an obstacle with respect to a captured image having a degree of similarity that is lower than a threshold value.
In the processing device according to the fifth aspect, the receiver unit receives the image information from a plurality of vehicles, the calculation unit calculates the degree of similarity of captured images, and the processing unit executes image processing on a captured image having a degree of similarity that is lower than a threshold value. As described above, the degree of similarity of the captured image is calculated as a value with respect to a reference image that has already been received. In the processing device, the configuration is such that a captured image having a degree of similarity that is equal to or higher than the threshold value is treated as an image that is similar to the reference image, and image processing is not executed. As a result, when performing image processing to identify an obstacle based on images captured at a vehicle, it is possible to reduce the number of captured images that are to be processed by the processing device and the processing load can be reduced.
A sixth aspect is a relay device that includes a receiver unit, a calculation unit, and a transmission unit. The receiver unit receives, from a vehicle, image information according to a captured image captured by an image capture device provided at the vehicle. The calculation unit calculates a degree of similarity between a captured image according to newly received image information, and a reference image according to image information that has already been received. The transmission unit transmits image information according to a degree of similarity that is lower than a threshold value to a processing device that performs image processing to identify an obstacle.
The relay device of the sixth aspect is a device that performs relay between a vehicle having an image capture device and a processing device that performs image processing to identify an obstacle. In the relay device, the receiver unit receives image information from a plurality of vehicles, the calculation unit calculates the degree of similarity of captured images, and the transmission unit transmits a captured image having a lower degree of similarity than a threshold value to the processing device. As described above, the degree of similarity of the captured image is calculated as a value with respect to a reference image that has already been received. According to this relay device, when performing image processing to identify an obstacle based on images captured at a vehicle, it is possible to reduce the number of captured images that are to be processed by the processing device and the processing load can be reduced.
A seventh aspect is a non-transitory recording medium storing a program that is executable by a computer to perform processing including receiver processing, calculation processing, and image processing. The receiver processing receives, from a vehicle, image information according to a captured image captured by an image. capture device provided at the vehicle. The calculation processing calculates a degree of similarity between a captured image according to newly received image information, and a reference image according to image information that has already been received. The image processing identifies an obstacle with respect to a captured image having a degree of similarity that is lower than a threshold value.
The program recorded on the non-transitory recording medium of the seventh aspect includes processing such that image information is received from a plurality of vehicles in a computer, the degree of similarity of captured images is calculated, and image processing is executed with respect to a captured image having a lower degree of similarity than a threshold value, As described above, the degree of similarity of the captured image is calculated as a value with respect to a reference image that has already been received. The program treats a captured image having a degree of similarity that is equal to or higher than a threshold value as an image that is similar to the reference image, and controls the computer so that image processing is not executed, As a result, when performing image processing to identify an obstacle based on images captured at a vehicle, it is possible to reduce the number of captured images that are to be processed by the computer that performs the image processing and the processing, load can be reduced.
According to the present disclosure, when performing image processing to identify an obstacle based on images captured at a vehicle, it is possible to reduce the number of captured images that are to be processed by the processing-side device and the processing load can be reduced.
Exemplary embodiments of the present disclosure will be described in detail based on the following figures, wherein:
As shown in
As shown in
In the present embodiment, the passing vehicle 12A corresponds to the “vehicle” of the present disclosure, and the peripheral vehicle 12B corresponds to the “other vehicle” of the present disclosure. The peripheral vehicle 12B is not limited to a following vehicle that travels in the same lane as the passing vehicle 12A on the traveling road T, but also includes an oncoming vehicle that travels in the oncoming lane and a vehicle that travels on another traveling road connected to the traveling road T.
The control device 20 of each vehicle 12 and the edge server 32 are connected to each other via a network N1. The edge server 32 and the center server 34 are connected to each other via a network N2.
In the image processing system 10 of
(Vehicle)
As shown in
The control device 20 is configured to include a CPU 20A, a ROM 20B, a RAM 20C, a communication I/F 20E, and an input/output I/F 20F. The CPU 20A, the ROM 20B, the RAM 20C, the communication I/F 20E, and the input/output I/F 20F are connected so as to be capable of communicating with each other through a bus 20G. CPU means Central Processing Unit, ROM means Read Only Memory, RAM means Random Access Memory, and I/F means interface.
The CPU 20A is a central computation processing unit that executes various programs and controls the respective units. Namely, the CPU 20A reads a program from the ROM 20B, and executes the program employing the RAM 20C as a workspace. The CPU 20A is an example of a first processor.
The ROM 20B stores various programs and various data. A control program 200 and image data 210 are stored in the ROM 20B of this embodiment. The control program 200 is a program for performing the image provision processing described below. The image data 210 stores image information regarding a captured image captured by the camera 22. The image data 210 may be stored in an SD card, a USB memory, or the like, connected to the control device 20. SD stands for Secure Digital and USB stands for Universal Serial Bus.
The RAM 20C serves as a workspace for temporary storage of programs and data.
The communication I/F 20E is an interface for connecting to the network N1 to communicate with the edge server 32. For communication with the edge server 32, a communication standard such as 5G, LTE, or Wi-Fi (registered trademark) is used, for example.
The input/output I/F 20F is an interface for communicating with the camera 22, the GPS receiver 24, and the notification device 26 mounted on the vehicle 12.
The camera 22 is fixed to the upper part of the front window of the vehicle 12, for example, and captures images to the front of the vehicle 12. Note that the camera 22 may not only capture images to the front of the vehicle 12 but may also capture images to be rear of the vehicle 12 and of the interior of the vehicle cabin.
The GPS receiver 24 receives GPS signals from plural GPS satellites, and measures the current position of the vehicle 12.
The notification device 26 is a device that is provided in the cabin of the vehicle 12 and that notifies the driver of information regarding the obstacle O. The notification device 26 may be a meter unit, a car navigation system, or an independent. device.
The acquisition unit 250 has a function of acquiring image information regarding a captured image captured by the camera 22, position information of the vehicle 12 detected by the GPS receiver 24, the current time, and the like. When the camera 22 images the obstacle O on the traveling path T, the acquisition unit 250 acquires image information triggered by an imaging instruction from the driver, position information, and information such as the current time. The imaging instruction corresponds, for example, to operation of a switch for imaging. In this embodiment, the position information of the passing vehicle 12A corresponds to the “first position information” of the present disclosure, and the position information of the surrounding vehicle 12B corresponds to the “second position information” of the present disclosure. The captured image acquired in the present embodiment may be a moving image or a still image.
The transmission unit 260 has a function of transmitting the image information acquired by the acquisition unit 250 to the edge server 32. Here, attachment information corresponding to the image information is added to the image information transmitted by the transmission unit 260. The additional information includes the position information of the vehicle 12 when the captured image was captured, the information of the traveling road T, the traveling direction, and the traveling time.
The notification receiving unit 270 has a function of receiving danger information (described below) transmitted from the center server 34. The danger information is notified from the center server 34 via the edge server 32 when the vehicle 12 is the surrounding vehicle 12B and approaches the obstacle O specified based on the captured image captured by the passing vehicle 12A.
The notification unit 280 has a function of notifying the driver that the obstacle O is approaching, based on the danger information received from the center server 34. The notification unit 280 operates the notification device 26 at the timing when the danger information is received.
(Server)
As illustrated in
The storage 30D is configured by a Hard Disk Drive or a Solid State Drive and stores various programs and various data.
The communication I/F 30E is an interface for communicating with the control device 20 and another server 30. The communication I/F 30E of the edge server 32 is an interface connectable to both the network N1 and the network N2. The network N2 may be a dedicated line or a line based on a communication standard such as 5G or LTE. The communication I/F 30E of the center server 34 is an interface connectable to at least the network N2.
The CPU 30A reads a program from the storage 30D and executes the program using the RAM 30C as a work area. In the CPU 30A of the present embodiment, different functions are realized in the edge server 32 and in the center server 34.
(Edge Server)
As shown in
The receiver unit 350 has a function of receiving the image information to which the attached information is added, which is transmitted from the plural control devices 20.
The calculation unit 352 has a function of calculating the degree of similarity between the captured image related to the newly received image information, and a reference image related to image information that has already been received. The image information regarding the reference image is stored in the image data 310 in the storage 30D. The reference image is the first image captured among the captured images acquired from the passing vehicle 12A at the capturing location of the captured image. Here, a known method is used to calculate the degree of similarity. For example, the feature points of the captured image and the feature points of the reference image can be respectively extracted, and the degree of similarity can be calculated from the degree of matching between the respective feature points.
The selectin unit 354 has a function of comparing the similarity degree calculated by the calculation unit 352 with a preset threshold value and selecting a captured image to be transmitted to the center server 34. Specifically, the selection unit 354 selects the image information related to a degree of similarity lower than the threshold value as “information to be transmitted to the center server 34” and provides the transmission unit 356 with the selected image information. Further, the selection unit. 354 selects the image information related to a degree of similarity equal to or higher than the threshold as “information to be discarded” and deletes this from the RAM 30C. Furthermore, when a reference image is newly set, the selection unit 354 stores the reference image in the image data 310.
The transmission unit 356 has a function of transmitting the image information selected by the selection unit 354 to the center server 34. Attachment information corresponding to the image information is added to the image information transmitted by the transmission unit 356.
(Center Server)
As shown in
The receiving unit 360 has a function of receiving the image information to which the attached information is added, which is transmitted from one or more edge servers 32.
The image processing unit 362, as a processing unit, has a function of performing image processing of identifying the obstacle O with respect to a captured image having a similarity degree lower than the threshold value. The image processing is executed by inputting image information to the learned model 330 that has performed machine-learning using previously known obstacles as teacher data. When it is possible to identify what the obstacle O is as a result of the image processing, information on the shape and size of the obstacle O, position information, and the like are stored as obstacle information in the obstacle database 340.
The determination unit 364 has a function of determining whether the surrounding vehicle 12B has approached the obstacle O. The determination unit 364 determines from the position information given to the obstacle O that the vehicle 12B has approached the obstacle O when the surrounding vehicle 12B is present at a predetermined distance. The predetermined distance is within the range of the distance R referred to in
The notification unit 366 has a function of notifying the surrounding vehicle 12B of the danger information. When the determination unit 364 determines that the obstacle O has been approached, the notification unit 366 of the present embodiment notifies the surrounding vehicle 12B of the danger information. That is, the center server 34 transmits the danger information to the control device 20 of the peripheral vehicle 12B via the edge server 32.
(Control Flow)
An example of the flow of processing executed in the image processing: system 10 of this embodiment will be described with reference to the flowcharts of
First, when the obstacle O has fallen on the traveling path T, the flow of the image provision processing executed in the control device 20 of the passing vehicle 12A that has passed by avoiding the obstacle O is as follows.
In step S100 of
In step S101, the CPU 20A adds to the captured image, as additional information, the position information of the vehicle 12, specifically, of the passing vehicle 12A, the information on the traveling road T, the traveling direction, and the traveling time.
In step S102, the CPU 20A transmits the image information with the attached information to the edge server 32. Then, the image providing process ends.
Next, the flow of the image selection processing executed in the edge server 32 is as follows.
In step S200 of
In step S201, the CPU 30A performs preprocessing on the captured image whose degree of similarity is to be calculated. Specifically, the CPU 30A performs area division and edge detection on the captured image.
In step S202, the CPU 30A calculates the degree of similarity, with respect to the reference image, for the captured image that has undergone the preprocessing.
In step S203, the CPU 30A determines whether or not the degree of similarity is low. Specifically, it is determined whether or not the calculated degree of similarity is below a preset threshold. In addition, when there is no reference image to be compared, that is, even for the first captured image of the obstacle O, the CPU 30A determines that the degree of similarity is low. When the CPU 30A determines that the degree of similarity is low, that is, when the degree of similarity is lower than the threshold value, the process proceeds to step S204. On the other hand, if the CPU 30A determines that the degree of similarity is not low, that is, if the degree of similarity is greater than or equal to the threshold value, the process proceeds to step S206.
In step S204, the CPU 30A stores, in the image data 310, image information related to the captured image that has a low degree of similarity. Here, if the captured image is the first image of the obstacle O captured, the captured image is stored in the image data 310 as a reference image.
In step S205, the CPU 30A transmits image information regarding the captured image, which has been determined to have a low degree of similarity, to the center server 34. Then, the image selecting process ends.
In step S206, the CPU 30A discards the captured image regarding which it has been determined that the degree of similarity is not low. That is, the CPU 30A erases the image information relating to the captured image from the RAM 20C without storing it in the image data 310. Then, the image selecting process ends.
The image selection processing executed by the edge server 32 ends when the obstacle information is received from the center server 34. This is because it is not necessary to select the captured image to be transmitted to the center server 34 in order to perform the image processing, because the obstacle O has been identified.
Next, the flow of the information processing executed in the center server 34 is as follows,
In step S300 of
In step S301, the CPU 30A executes image processing. That is, the CPU 30A inputs the image information to the learned model 330 and tries to identify the obstacle O.
In step S302, the CPU 30A determines whether or not the obstacle O has been identified. When the CPU 30A determines that the obstacle O has been specified, the processing proceeds to step S303. On the other hand, when the CPU 30A determines that the obstacle O has not been specified, the CPU 30A ends the information processing.
In step S303, the CPU 30A stores the obstacle information. Specifically, the CPU 30A stores, as the obstacle information, information on the type and size of the identified obstacle O and position information in the obstacle database 340. Then, the information processing ends.
On the other hand, the flow of the notification processing executed in the center server 34 after the obstacle O has been specified is as follows.
In step S310 of
In step S311, CPU 30A determines whether or not peripheral vehicle 12B has approached Obstacle O. For example, as shown in
In step S312, the CPU 30A notifies the surrounding vehicle 12B that has approached the obstacle O of the obstacle information as the danger information. The notification processing is then ended. Although the center server 34 of the present embodiment transmits the obstacle information to the control device 20 of the peripheral vehicle 12B via the edge server 32, the obstacle information may be transmitted to the control device 20 directly without passing through the edge server 32.
Through the above notification processing, the peripheral vehicle 12B that has received the obstacle information alerts the driver to the obstacle O.
(Summary)
The image processing system 10 of the present embodiment includes a control device 20 mounted on the vehicle 12, a center server 34 that executes image processing, and an edge server 32 that relays between the control device 20 and the center server 34. In the present embodiment, when the passing vehicle 12A images the obstacle O on the traveling path T, the edge server 32 receives image information from the control devices 20 of plural passing vehicles 12A, and the calculating unit 352 of the edge server 32 calculates the degree of similarity of the captured images. Then in the center server 34, the image processing unit 362 performs the image processing on a captured image having a lower degree of similarity than the threshold, to identify the obstacle O.
Here, the degree of similarity of the captured image is calculated as a value with respect to the reference image. Then, the edge server 32 is configured so that a captured image having a degree of similarity that is equal to or higher than a threshold is treated as an image that is similar to the reference image and is not transmitted to the center server 34. That is, the image similar to the reference image is not received by the center server 34, and the image processing is not executed.
As described above, according to the present embodiment, when the image processing for identifying the obstacle O is performed based on the image captured by the vehicle 12, the captured images to be processed by the center server 34 can be reduced, and the processing load can be reduced.
Further, according to the present embodiment, since the edge server 32 selects the captured image to be transmitted to the center server 34 based on the degree of similarity, the amount of communication in the network N2 between the edge server 32 and the center server 34 can be reduced.
Further, according to the present embodiment, since an image that is determined to be similar to the reference image in the edge server 32 is discarded and is not stored in the storage 30D, the resource consumption of the edge server 32 can be suppressed.
Furthermore, according to the present embodiment, it is possible to call the attention of the peripheral vehicle 12B to the obstacle O on the traveling path T based on the position information of the obstacle O.
In the first embodiment, the edge server 32 is configured to execute the image selection processing, but in the second embodiment, the edge server 32 is eliminated and the center server 34 can execute the image selection processing. Explanation follows regarding the points that differ from the first exemplary embodiment. Configurations corresponding to those of the first exemplary embodiment are allocated the same reference numerals, and explanation thereof is omitted.
The image processing system 10 of the present embodiment includes a plurality of vehicles 12, a control device 20 mounted on each vehicle 12, and a center server 34.
In the center server 34 of the present embodiment, when the image information is received from the control device 20, the image selection processing is executed first, and then the information processing is executed. Then, after the obstacle O is specified, the notification processing is executed.
In the present embodiment, similarly to the first embodiment, when the image processing for identifying the obstacle O is performed based on the image captured by the vehicle 12, the captured images to be processed by the center server 34 can be reduced, and the processing load can be reduced.
In each of the above-described embodiments, among the captured images acquired from the passing vehicle 12A, the image captured first at the capturing location of the captured images is used as the reference image, hut the present disclosure is not limited to this. For example, the calculation unit 352 may use, as the reference image, the captured image with the highest image quality among the images captured by the camera 22 of the passing vehicle 12A. Further, for example, the captured image captured by the camera 22 having the highest resolution among the respective passing vehicles 12A may be used as the reference image.
Further, when the calculation unit 352 receives a captured image having a higher image quality than the reference image to be compared when calculating the degree of similarity, the calculation unit 352 may update the captured image as a new reference image.
It should be noted that various processors other than a CPU may execute the various processes executed by the CPUs 20A and 30A by reading the software including the program in each of the above embodiments. Examples of such processors include programmable logic devices (PLD) with circuit configurations that are reconfigurable after manufacture, such as field-programmable gate arrays (FPGA), and dedicated electronic circuits that are processors including circuit configurations custom designed to execute specific processing, such as application specific integrated circuits (ASIC) or the like. Moreover, the above-described receiver processing may be executed by one of such processors, or may be executed by a combination of two or more processors of the same type or different types (for example, plural FPGAs, or a combination of a CPU and an FPGA). More specific examples of hardware structures of such processors include electric circuits configured by combining circuit elements, such as semiconductor devices.
In each of the exemplary embodiments described above, explanation has been given in which the respective programs are provided in a format pre-stored or installed on a non-transitory computer-readable recording medium. For example, in the control device 20, the control program 200 is stored in the ROM 20B in advance. Further, example, the relay program 300 in the edge server 32 and the processing program 320 in the center server 34 are stored in advance in the storage 30D of each server 30. However, there is no limitation thereof, and the respective programs may be provided in a format stored on a non-transitory recording medium such as Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc Read Only Memory (DVD-ROM), or Universal Serial Bus (USB) memory. The programs may alternatively be provided in a format that can be downloaded from an external device through a network.
The processing flows described in the above exemplary embodiment are merely examples thereof, and unnecessary steps may be omitted, new steps added. or the processing sequence changed within a range not departing from the spirit thereof.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2019-208883 | Nov 2019 | JP | national |
