Information processor, information processing method, and information processing system
The information processing device detects vehicles by analyzing high-brightness areas in images, overcoming low-light challenges by dividing and counting bright pixels, enhancing detection accuracy in adverse conditions.
Patent Information
- Application Number
- JP2024023942
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-02-20
- Publication Date
- 2025-09-01
AI Technical Summary
Existing image-based vehicle detection systems struggle to accurately identify vehicles in low-light or wet conditions due to difficulty in recognizing feature points, such as vehicle outlines.
An information processing device and method that utilizes a judgment unit to detect high-brightness areas in images, dividing the image into sections, calculating the proportion of bright pixels, and determining the presence of vehicles based on consecutive sections with high brightness, allowing detection even when feature points are unclear.
Enables accurate vehicle detection in challenging conditions by utilizing high-brightness areas, reducing false positives from non-vehicle light sources, and improving detection accuracy.
Smart Images

Figure 2025127288000001_ABST
Abstract
Description
[Technical Field]
[0001] The present invention relates to an information processing device, an information processing method, and an information processing system. [Background technology]
[0002] Technologies have been developed that capture images of a predetermined area, including a roadway, and analyze the resulting images to detect the presence or absence of vehicles within the predetermined area. For example, the device described in Patent Document 1 captures images of a predetermined, approximately rectangular area set near a road intersection with a camera, and detects vehicles by performing image processing such as edge extraction on the resulting images. The results of vehicle detection are used to adjust the lighting times of traffic lights, thereby contributing to easing congestion near intersections. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-251910 Summary of the Invention [Problem to be solved by the invention]
[0004] However, the above-mentioned technology does not function accurately when it is difficult to recognize feature points such as the outline of a vehicle in the obtained image. For example, when taking an image in an environment where the surroundings of the camera are not bright enough, such as at night, and the road surface is wet, it may not be possible to detect the presence or absence of a vehicle within a specified area.
[0005] An object of the present invention is to provide an information processing device, an information processing method, and an information processing system that can detect the presence or absence of a vehicle within a predetermined area even in situations where it is difficult to find the vehicle's feature points on an image. [Means for solving the problem]
[0006] In order to solve the above problems and achieve the objectives, an information processing device that is one aspect of the present invention is characterized by comprising a judgment unit that judges whether or not a high-brightness area that meets a predetermined criterion is present in a detection target image obtained by photographing a predetermined area including a vehicle's driving path, and a detection unit that detects the presence of a detection target vehicle when the judgment unit judges that the high-brightness area is present.
[0007] Here, the detection target image can be configured from an image obtained by photographing the roadway from above and in front along the roadway.
[0008] The determination unit may also include a division unit that divides the detection target image into a plurality of sections lined up along the roadway, a calculation unit that calculates the proportion of pixels having a predetermined brightness or higher for each section, a counting unit that measures the number of consecutive sections along the roadway where the proportion is equal to or higher than the predetermined value, and a determination unit that determines whether or not there is a high-brightness area based on the number of consecutive sections.
[0009] Furthermore, the determination unit can determine that the high brightness area exists when the number of consecutive occurrences is three or more.
[0010] Furthermore, the lengths from the front to the rear of each of the plurality of sections may be equal when converted into actual distances on the travel path.
[0011] Furthermore, an information processing method that is one aspect of the present invention is an information processing method executed by a computer to detect a vehicle, and is characterized by comprising a judgment process that judges whether or not a high-brightness area that meets a predetermined criterion is present in a detection target image obtained by photographing an area including the vehicle's driving path, and a detection process that detects the presence of the detection target vehicle when the judgment process determines that the high-brightness area is present.
[0012] Furthermore, an information processing system that is one aspect of the present invention is an information processing system that includes an imaging device that acquires a detection target image that captures a specified area including a vehicle's driving path, and an information processing device that detects the vehicle based on the detection target image, wherein the information processing device is characterized by including a determination unit that determines whether or not a high-brightness area that meets specified criteria is present in the detection target image, and a detection unit that detects the presence of the detection target vehicle when the determination unit determines that the high-brightness area is present. [Effects of the Invention]
[0013] According to the present invention, it is possible to provide an information processing device, an information processing method, and an information processing system that detect the presence or absence of a vehicle within a predetermined area even in a situation where it is difficult to find the characteristic points of the vehicle on the image. [Brief explanation of the drawings]
[0014] [Figure 1] 1 is a block diagram showing a configuration of an information processing system according to an embodiment of the present invention; [Figure 2] 10A and 10B are diagrams illustrating an example of a detection target image determined by a determination unit of the information processing system. [Figure 3] FIG. 2 is a block diagram showing the configuration of a determination unit. [Figure 4] 10 is a flowchart of the operation of the information processing device. [Figure 5] 10 is a flowchart showing details of a determination process. [Figure 6] FIG. 10 is a diagram showing a detection target image after division processing has been performed. [Figure 7] FIG. 10 is a diagram showing the lengths of sections generated by the division process. [Figure 8] FIG. 10 is a diagram showing a table summarizing the results of an example of a determination process. DETAILED DESCRIPTION OF THE INVENTION
[0015] An embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing the configuration of an information processing system 1 according to this embodiment. The information processing system 1 is one of systems that monitors roads using cameras or the like, and is used, for example, to detect the presence or absence of a vehicle 113 (see Fig. 2) on a road 112.
[0016] As shown in FIG. 1, the information processing system 1 includes an imaging device 100 and an information processing device 200.
[0017] The imaging device 100 includes an imaging unit 110 and a communication unit 120. The imaging unit 110 is configured with a device such as a camera, and captures an image (still image or video) including a road 112 on which a vehicle 113 is traveling. By capturing an image using the imaging unit 110, a detection target image 111 is acquired, which is an image of a predetermined area for determining the presence or absence of a vehicle 113, as shown in a solid-line frame in FIG. 2 .
[0018] In this embodiment, the imaging unit 110 is installed above the roadway 112, faces the opposite direction to the traveling direction so as to capture the area ahead of the traveling vehicle 113, and is configured to capture an image of the roadway 112 from above and in front along the roadway 112. Therefore, the detection target image 111 is an image obtained by capturing an image of the roadway 112 from above and in front along the roadway 112.
[0019] 1, the communication unit 120 is connected to the information processing device 200 via a network N such as the Internet, and transmits data of the detection target image 111 based on a command from the information processing device 200. Note that the communication unit 120 may be connected to the information processing device 200 wirelessly or by wire. Also, the communication unit 120 may be omitted and the imaging device 100 may be incorporated into the information processing device 200 to form a single device.
[0020] The information processing device 200 is configured by a computer such as a server, and detects the presence or absence of a vehicle 113 on a road 112 based on data of an acquired detection target image 111. The information processing device 200 includes a communication unit 210, a storage unit 220, and a control unit 230.
[0021] The communication unit 210 is connected to the imaging device 100 via the above-mentioned network N, and receives data of the detection target image 111 transmitted from the communication unit 120 of the imaging device 100. The storage unit 220 is configured with a storage device such as a hard disk or memory, and stores programs necessary for the operation of the information processing system 1, as well as the data of the detection target image 111 received by the communication unit 120.
[0022] The control unit 230 is configured with a calculation device such as a CPU (Central Processing Unit), and includes an acquisition unit 240, an extraction unit 250, a determination unit 260, and a detection unit 270. The acquisition unit 240 operates the communication unit 210 to acquire data of the detection target image 111 from the imaging device 100 at predetermined time intervals and temporarily store the data in the storage unit 220. Note that the predetermined time can be set appropriately, and by setting the data acquisition interval to be every few seconds, images may be acquired almost in real time, or by setting the data acquisition interval to be longer than every few seconds, the amount of acquired data may be reduced.
[0023] The extraction unit 250 analyzes the detection target image 111 and extracts feature points of the vehicle 113 on the detection target image 111. The feature points may be, for example, the contour lines of each part of the vehicle 113. In this embodiment, since the contour lines of parts of the two vehicles 113 located on the front side of the road 112 can be recognized in the detection target image 111 shown in FIG. 2, these contour lines may be extracted as feature points. In this way, the extraction unit 250 mainly operates when at least parts of the vehicles 113 can be recognized on the detection target image 111.
[0024] On the other hand, the determination unit 260 mainly operates when the vehicle 113 cannot be directly recognized on the detection target image 111. A case in which the vehicle 113 cannot be directly recognized is, for example, a case in which so-called halation H (shown by a point cloud in FIG. 2 ), which is a blown-out highlight caused by pixel overflow, occurs on the detection target image 111. When the information processing system 1 is used, for example, at night or in rainy weather, the sensitivity of a sensor (not shown) in the imaging unit 110 is often increased to amplify the amount of detected light in order to make the vehicle 113 easier to recognize. For this reason, the above-mentioned halation H is likely to occur over a wide area due to light from vehicle headlights reflected on the road surface of the road 112, and the vehicle 113 may be hidden by the halation H and may not be directly recognized. Therefore, the determination unit 260 operates when the vehicle 113 cannot be directly recognized due to the occurrence of such halation H.
[0025] The determination unit 260 analyzes the detection target image 111 and determines whether or not a high-brightness area h (see FIG. 6) that satisfies a predetermined criterion, which will be described later, exists in the detection target image 111. As shown in FIG. 3, the determination unit 260 includes a division unit 261, a calculation unit 262, a counting unit 263, and a determination unit 264. Details of the operation of the information processing device 200, including the operation of the determination unit 260, will be described later.
[0026] The detection unit 270 shown in Fig. 1 detects the presence or absence of a vehicle 113 based on the feature points extracted by the extraction unit 250. Furthermore, when the determination unit 260 determines that "a high-brightness area h is present," the detection unit 270 detects the presence of a detection target vehicle 114 (see Fig. 6). The "detection target vehicle 114" is "a vehicle 113 that emitted light from a vehicle headlight or the like that generated the high-brightness area h," and also "a vehicle 113 that is traveling or is about to travel forward on the road 112 toward the high-brightness area h."
[0027] With the above configuration, the information processing system 1 detects the presence or absence of the vehicle 113 (and the detection target vehicle 114) based on the feature points of the vehicle 113 extracted on the detection target image 111, or the high-brightness area h whose presence or absence has been determined on the detection target image 111.
[0028] Next, the operation of the information processing system 1, that is, an information processing method for detecting the vehicle 113, will be described. The information processing method is implemented by a computer such as the information processing device 200. First, an acquisition process is executed as shown in FIG. 4 (step S1). In step S1, the communication unit 210 receives a command from the acquisition unit 240 of the control unit 230 and operates to acquire data of the detection target image 111 from the imaging device 100 at predetermined time intervals. The acquired data is temporarily stored in the storage unit 220.
[0029] Next, extraction processing is executed (step S2). In step S2, the extraction unit 250 of the control unit 230 analyzes the data of the detection target image 111 and extracts feature points of the vehicle 113.
[0030] As described above, the feature points may be the contour lines of each part of the vehicle 113. Note that this extraction process may be set to operate during times when the contour lines of each part of the vehicle 113 are easy to recognize, such as daytime, and may be omitted at night or during rainy weather. Furthermore, a predetermined process (not shown) may be executed before the extraction process to determine whether or not the contour lines of the vehicle 113 can be extracted, and whether or not to execute the extraction process may be determined based on the results of that process.
[0031] Next, a determination process is executed (step S3). In step S3, the determination unit 260 of the control unit 230 determines whether or not a high-brightness region h that satisfies a predetermined criterion is present on the detection target image 111. In this step S3, first, a division process is executed (step S31), as shown in FIG. 5. In step S31, the division unit 261 divides the detection target image 111 into n pieces (five in this embodiment) of predetermined dimensions along the traveling direction of the road 112, as shown in FIG. 6. As a result, a plurality of sections E1, E2, E3, E4, and E5 lined up along the road 112 are generated on the detection target image 111.
[0032] In this case, as described above, the area from which the feature points of the vehicle 113 can be extracted by the extraction unit 250 does not need to be subject to division. For example, in the detection target image 111 in Fig. 6, parts of the two vehicles 113 can be recognized in the front area, and feature points can be extracted from these parts, so this part is not divided.
[0033] 7, it is preferable to set the lengths e1, e2, e3, e4, and e5 from front to rear of each of the sections E1, E2, E3, E4, and E5 so that they are equal when replaced with the actual distance on the road 112. It is even more preferable to set the lengths e1, e2, e3, e4, and e5 to be approximately equal to the total length of one vehicle 113, for example. This makes it possible to set the sections E1, E2, E3, E4, and E5 in accordance with the sections obtained by equally dividing the actual road 112 at distances approximately equal to the total length of one vehicle 113.
[0034] 5, initialization is performed (step S32). In step S32, the counting unit 263 sets a value called the determination number T to 0. The "determination number T" is defined as "the number of partitions E1, E2, E3, E4, and E5 for which the calculation process described below has already been completed."
[0035] Next, a measurement process is executed (step S33). In step S33, the counting unit 263 counts the number of determinations T for the sections E1, E2, E3, E4, and E5. If the result of the measurement is that the number of determinations T is N, the determination unit 264 determines that there is "no high brightness region h" for the high brightness region h described below, and the determination process is completed.
[0036] On the other hand, if the number of determinations T is less than N, the process proceeds to step S34. The value of N is set according to the number of the above-mentioned sections E1, E2, E3, E4, and E5. In this embodiment, since sections E1 to E5 are generated by the division process, N is set to 5.
[0037] In step S34, a calculation process is executed. In this step S34, the calculation unit 262 calculates the proportion of pixels having a predetermined brightness or more for one of the sections E1, E2, E3, E4, and E5 for which the calculation process has not yet been completed. For example, in this embodiment, among the sections E1, E2, E3, E4, and E5, the proportion of "pixels having a predetermined brightness or more" is first calculated for section E1.
[0038] For example, if the brightness range is 0 to 255, the "pixels with a brightness value of 240 or more" are "pixels with a brightness value of 240 or more." The percentage of pixels with a brightness value of 240 or more in a partition is calculated by dividing the number of pixels in the partition with a brightness value of 240 or more by the total number of pixels in the partition × 100. If the image captured by the imaging unit 110 is a color image, it is recommended to convert it to a grayscale image before carrying out the above calculation process.
[0039] Next, a counting process is executed (step S35). In step S35, the ratio of "pixels having a predetermined brightness or more" calculated by the calculation process is confirmed by the counting unit 263. Then, a value called the consecutive number t is measured. The "consecutive number t" is defined as "the number of consecutive sections E1, E2, E3, E4, and E5 along the travel path 112 where the ratio of pixels having a predetermined brightness or more (in this embodiment, a brightness value of 240 or more) is a predetermined value or more (in this embodiment, 90% or more)."
[0040] Specifically, in step S35, if the result of checking the proportion of "pixels having a predetermined brightness or more" is that the proportion is less than 90%, the process proceeds to step S351. In step S351, the counting unit 263 executes a process of adding 0 to the value of the consecutive number t. Then, the process returns to step S33. On the other hand, if the result of checking the proportion of "pixels having a predetermined brightness or more" is that the proportion is 90% or more, the process proceeds to step S352. In step S352, the counting unit 263 executes a process of adding 1 to the value of the consecutive number t. Then, the process proceeds to step S36.
[0041] In step S36, a determination process is performed. In the determination process, the determination unit 264 determines whether or not there is a high-luminance region h, which is indicated by a solid-line frame in FIGS. 6 and 7, based on the consecutive number t counted by the counting unit 263. In this embodiment, if the consecutive number t is 3 or more, it is determined that there is a high-luminance region h, and the determination process is completed. On the other hand, if the consecutive number t is less than 3, it is not determined that there is a high-luminance region h, and the process returns to step S33. The "high-luminance region h" refers to "a region where halation H is occurring, which particularly suggests the presence of the detection target vehicle 114."
[0042] When a detection target vehicle 114 is present, halation H appears along the roadway 112 due to the vehicle headlights reflected on the road surface of the roadway 112, and the halation H is detected continuously from the distance where the detection target vehicle 114 is traveling toward the front where the imaging unit 110 is installed. Therefore, in order to determine that such a characteristic halation H portion is a high brightness region h, in this embodiment, when the above-mentioned consecutive number t is 3 or more, it is determined that "a high brightness region h exists."
[0043] In this way, steps S33 to S36 are looped until processing is completed for all sections E1 to E5 and it is determined in step S33 that "high luminance area h is not present," or until it is determined in step S36 that "high luminance area h is present."
[0044] Alternatively, as shown in Fig. 8, measurement and counting processes may be performed on all generated sections E1, E2, E3, E4, and E5, regardless of whether or not a high-brightness area h is present, and then a determination process may be performed. Fig. 8 is a table summarizing the results of an example of the determination process.
[0045] 8, the left column lists the codes of the sections, the center column lists the results of the calculation process, and the right column lists the results of the counting process. In the counting process results, an × is listed when the process of adding 0 to the consecutive number t is performed in step S351. On the other hand, a ◯ is listed when the process of adding 1 to the consecutive number t is performed in step S352.
[0046] If a determination process is then performed based on this result, the consecutive number t for sections E1 to E3 is 3, and so sections E1 to E3 are determined to have a high-brightness area h. On the other hand, the consecutive number t for sections E4 to E5 is less than 3, and so they are determined to have no high-brightness area h.
[0047] Finally, as shown in Fig. 4, the detection process is executed (step S4). In step S4, if the detection unit 270 determines that "high luminance area h exists," it detects that the detection target vehicle 114 exists. This completes the information processing method.
[0048] Information about the vehicle 113 (detection target vehicle 114) detected by the information processing method can be used for various purposes. For example, the information processing system 1 may be installed near an intersection with a traffic light, and information about the detected vehicle 113 (detection target vehicle 114) may be used to control the traffic light. Specifically, when the information processing method detects that the vehicle 113 is present, the lighting time of the green light may be extended, and when the information processing method detects that the vehicle 113 is absent, the lighting time of the red light may be extended.
[0049] As described above, according to the embodiment described above, the presence or absence of the detection target vehicle 114 (vehicle 113) can be detected by detecting the presence or absence of the high-brightness region h, and it is not necessary to find feature points such as the outline of the vehicle 113 in the detection target image 111 when detecting the vehicle 113. Therefore, even if halation H occurs in the detection target image 111 due to light from vehicle headlights reflected on the road surface of the road 112, for example, the presence or absence of the vehicle 113 can be detected by utilizing the halation H. Therefore, it is possible to provide an information processing device 200 that detects the presence or absence of the vehicle 113 in a predetermined region even in a situation where it is difficult to find feature points of the vehicle 113 in the detection target image 111 (on the image).
[0050] Furthermore, according to this embodiment, the dividing unit 261, the calculating unit 262, and the counting unit 263 operate to calculate the number t of consecutive sections E1, E2, E3, E4, and E5 that have a high proportion of pixels with a predetermined brightness or higher on the detection target image 111. Then, the determining unit 264 operates to determine whether or not there is a high brightness region h based on the result of calculating the number t of consecutive sections.
[0051] Furthermore, according to this embodiment, when the continuation number t is less than 3, it is possible to prevent the corresponding sections E1, E2, E3, E4, and E5 from being determined as high-brightness regions h. This makes it possible to prevent pinpoint areas where brightness increases due to factors other than light emitted by the vehicle 113, such as around streetlights, from being erroneously detected as high-brightness regions h. This makes it possible to narrow down the high-brightness regions h to areas where the vehicle 113 is likely to exist, which contributes to improving the detection accuracy of the target vehicle 114 using the high-brightness regions h.
[0052] Furthermore, according to this embodiment, the lengths e1, e2, e3, e4, and e5 from the front to the rear of the sections E1, E2, E3, E4, and E5 generated by the division process can be set to be equal when replaced with the actual distance on the road 112. This makes it possible to detect the presence or absence of a vehicle 113 in the sections E1, E2, E3, E4, and E5 that correspond to the sections when the actual road 112 is evenly divided.
[0053] Furthermore, according to this embodiment, by operating the information processing device 200, an information processing method can be executed to detect the presence or absence of a vehicle 113 within a specified area even in a situation where it is difficult to find the feature points of the vehicle 113 from the detection target image 111 (on the image).
[0054] Furthermore, according to this embodiment, the imaging device 100 and the information processing device 200 can provide an information processing system 1 that detects the presence or absence of a vehicle 113 within a specified area even in a situation where it is difficult to find feature points of the vehicle 113 from the detection target image 111 (on the image).
[0055] The above-described embodiment merely illustrates one aspect of the information processing system 1, the information processing device 200, and the information processing method, and the aspect is not limited thereto. For example, in this embodiment, five partitions, E1, E2, E3, E4, and E5, are generated by the division process, but the number of partitions generated may be less than five or more than five. Furthermore, the lengths e1, e2, e3, e4, and e5 from the front to the rear of the partitions E1, E2, E3, E4, and E5 are not limited to being set so that they are equal when converted to actual distances, and may be set as appropriate.
[0056] Furthermore, when calculating the percentage of pixels with a predetermined brightness or more by the calculation process, "above a predetermined brightness" may be a value other than the brightness value 240. Furthermore, when the determination unit 264 determines whether or not there is a high brightness region h, the consecutive number t may be a value other than 3. However, as described above, from the viewpoint of avoiding erroneous detection of pinpoint areas with high brightness due to factors other than light emitted by the vehicle 113 as high brightness regions h, it is preferable to set the consecutive number t to 1 or more. [Explanation of symbols]
[0057] E1, E2, E3, E4, E5 compartments e1, e2, e3, e4, e5 length h High brightness area t consecutive number 1. Information Processing Systems 100 Imaging device 111 Detection target image 112 Running Track 113 vehicles 114 Vehicles to be detected 200 Information processing device 260 Judgment section 261 Division 262 Calculation Unit 263 Counting Unit 264 Decision Unit 270 Detection unit
Claims
1. a determination unit that determines whether or not there is a high-brightness area that satisfies a predetermined criterion in a detection target image obtained by capturing an image of a predetermined area including a vehicle travel path; a detection unit that detects the presence of a target vehicle when the determination unit determines that the high-brightness area exists.
2. 2. The information processing device according to claim 1, wherein the detection target image is an image obtained by photographing the road from above and in front along the road.
3. The determination unit a dividing unit that divides the detection target image into a plurality of sections that are arranged along the road; a calculation unit that calculates a ratio of pixels having a luminance equal to or greater than the predetermined luminance for each of the sections; a counting unit that counts the number of consecutive sections along the travel path where the ratio is equal to or greater than a predetermined value; The information processing apparatus according to claim 2 , further comprising: a determination unit that determines whether or not there is a high-luminance area based on the number of consecutive areas.
4. The information processing apparatus according to claim 3 , wherein the determination unit determines that the high-luminance area exists when the number of consecutive pixels is three or more.
5. 5. The information processing device according to claim 4, wherein the lengths from the front to the rear of each of the plurality of sections are equal when converted into actual distances on the road.
6. 1. A computer-implemented information processing method for detecting a vehicle, comprising: a determination process for determining whether or not there is a high-brightness area that satisfies a predetermined criterion in a detection target image obtained by photographing an area including the vehicle's running path; a detection process for detecting the presence of a target vehicle when the determination process determines that the high-brightness area exists.
7. an imaging device that captures a detection target image of a predetermined area including a vehicle travel path; an information processing device that detects the vehicle based on the detection target image, The information processing system is characterized in that the information processing device includes a judgment unit that judges whether or not there is a high-brightness area that meets a predetermined criterion on the detection target image, and a detection unit that detects the presence of a detection target vehicle when the judgment unit determines that the high-brightness area exists.
Citation Information
Patent Citations
Method and device for detecting preceding jam, and preceding jam sensitive control system
JP2006251910A