Image-based object detection method, object detection device, and traffic management system

Abstract

Embodiments of the present disclosure relate to an image-based object detection method, an object detection device, and a traffic management system for detecting, on the basis of an image of a road, object characteristic information of an object detected from the image.

Description

Image-based object detection method, object detection device, and traffic management system

[0001] Embodiments of the present disclosure relate to an image-based object detection method, an object detection device, and a traffic management system.

[0002] Currently, in the transportation industry, objects such as vehicles are detected based on video. However, this detection technology may not be able to accurately detect objects depending on factors such as video transmission conditions (network conditions) and the timing of video acquisition. In particular, if the object is moving, detection may be impossible or performance may be significantly reduced.

[0003] Embodiments of the present disclosure can provide an image-based object detection method, an object detection device, and a traffic management system capable of detecting objects more accurately.

[0004] Embodiments of the present disclosure can provide an image-based object detection method, an object detection device, and a traffic management system that can more accurately detect objects using machine learning algorithms such as deep learning.

[0005] Embodiments of the present disclosure can provide an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting object characteristic information representing characteristics of an object or characteristics of some components included in an object.

[0006] Embodiments of the present disclosure can provide an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting a moving object and accurately detecting object characteristic information about the moving object.

[0007] A traffic management system according to embodiments of the present disclosure may include an image acquisition device that acquires an image of a road and an object detection device that detects object characteristic information about an object detected from the image based on the image.

[0008] The object detection device detects an object in an image through deep learning, performs a first cropping process on a portion of the object in a current frame, extracts a first edge degree from a first object image of the first object obtained through the first cropping process, performs a second cropping process on a portion of the object in a next frame, extracts a second edge degree from a second object image of the first object obtained through the second cropping process, compares the first edge degree and the second edge degree, and if the first edge degree is stronger than the second edge degree, detects object characteristic information based on the first object image, and if the second edge degree is stronger than the first edge degree, detects object characteristic information based on the second object image.

[0009] In a traffic management system according to embodiments of the present disclosure, an object may be a moving mobile object.

[0010] In a traffic management system according to embodiments of the present disclosure, the image acquisition device and the object detection device can communicate in a wired or wireless manner.

[0011] In a traffic management system according to embodiments of the present disclosure, an object is a vehicle, object characteristic information includes vehicle characteristic information of the vehicle, and the vehicle characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0012] In a traffic management system according to embodiments of the present disclosure, an object is a pedestrian, object characteristic information includes pedestrian characteristic information of the pedestrian, and the pedestrian characteristic information may include information on one or more of the pedestrian's age, gender, walking status, clothing, and items worn.

[0013] An object detection device according to embodiments of the present disclosure may include an image storage unit that stores an image of a road, an object detection unit that detects an object through a machine learning algorithm based on the image, a frame selection unit that selects a first frame from among a plurality of frames included in the image, and an object characteristic detection unit that extracts an edge of an object from the first frame and detects object characteristic information about the object based on the extracted edge.

[0014] In the object detection device according to embodiments of the present disclosure, the object may be a moving object.

[0015] An object detection method according to embodiments of the present disclosure may include an image acquisition step of acquiring an image of a road through an image acquisition device, an object detection step of detecting an object through a machine learning algorithm based on the image by an object detection device, a frame selection step of selecting a first frame from among a plurality of frames included in the image by the object detection device, and an object characteristic detection step of extracting an edge of an object from the first frame by the object detection device and detecting object characteristic information about the object based on the extracted edge.

[0016] In the object detection step, the object detection device can detect an object including at least one of a vehicle, a pedestrian, and a road structure by executing deep learning as a machine learning algorithm based on an image.

[0017] In the frame selection step, the object detection device may select a first frame having the highest sharpness information among the plurality of frames based on sharpness information about the edge of the object extracted from each of the plurality of frames. Here, each of the plurality of frames may be an object image obtained by cropping a portion of the frame in which the object is displayed.

[0018] In the object detection method according to embodiments of the present disclosure, the object may be a moving object.

[0019] The object is a vehicle, and the object characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0020] The object is a pedestrian, and the object characteristic information may include information about one or more of the pedestrian's age, gender, walking status, clothing, and items worn.

[0021] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of detecting objects more accurately can be provided.

[0022] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of more accurately detecting objects using a machine learning algorithm such as deep learning can be provided.

[0023] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting object characteristic information representing characteristics of an object or characteristics of some components included in an object can be provided.

[0024] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting a moving object and accurately detecting object characteristic information about the moving object can be provided.

[0025] FIG. 1 is a block diagram of a traffic management system according to embodiments of the present disclosure.

[0026] FIG. 2 exemplarily illustrates an environment in which a traffic management system according to embodiments of the present disclosure is installed.

[0027] FIG. 3 illustrates an image for object characteristic detection according to embodiments of the present disclosure.

[0028] FIG. 4 illustrates multiple frames for object characteristic detection according to embodiments of the present disclosure.

[0029] FIG. 5 is a block diagram of an object detection device according to embodiments of the present disclosure.

[0030] FIG. 6 is a flowchart of an object detection method according to embodiments of the present disclosure.

[0031] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to exemplary drawings. When adding reference numerals to components in each drawing, identical components may have the same numerals as much as possible even if they are shown in different drawings. In addition, when describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description may be omitted. When "includes," "has," "consists of," etc. are used in this specification, other parts may be added unless "only" is used. When a component is expressed in the singular, it may include a plurality unless there is a special explicit description.

[0032] Additionally, terms such as first, second, A, B, (a), (b), etc. may be used to describe components of the present disclosure. These terms are only intended to distinguish the components from other components, and the nature, order, sequence, or number of the components are not limited by the terms.

[0033] In a description of the positional relationship of components, when it is described that two or more components are "connected," "combined," or "connected," it should be understood that the two or more components may be directly "connected," "combined," or "connected," but that the two or more components may also be further "interposed" with another component to be "connected," "combined," or "connected." Here, the other component may be included in one or more of the two or more components that are "connected," "combined," or "connected" to each other.

[0034] In the description of the temporal flow relationship related to components, operation methods, or manufacturing methods, for example, when the temporal or flow relationship is described as “after”, “following”, “next to”, “before”, etc., it may also include cases where it is not continuous, unless “immediately” or “directly” is used.

[0035] Meanwhile, when numerical values ​​or corresponding information (e.g., levels, etc.) for components are mentioned, even without separate explicit description, the numerical values ​​or corresponding information may be interpreted as including an error range that may occur due to various factors (e.g., process factors, internal or external impact, noise, etc.).

[0036] Hereinafter, various embodiments of the present disclosure will be described in detail with reference to the attached drawings.

[0037] FIG. 1 is a block diagram of a traffic management system (100) according to embodiments of the present disclosure. FIG. 2 exemplarily illustrates an environment in which a traffic management system (100) according to embodiments of the present disclosure is installed.

[0038] Referring to FIG. 1, a traffic management system (100) according to embodiments of the present disclosure may include an image acquisition device (110) that acquires an image of a road and an object detection device (120) that detects object characteristic information about an object detected from the image based on the image.

[0039] In embodiments of the present disclosure, object characteristic information may be information indicating characteristics of an object or characteristics of some configuration included in an object.

[0040] The object detection device (120) detects an object in an image through deep learning, performs a first cropping process on a portion of the object in a current frame, extracts a first edge degree from a first object image of the first object obtained through the first cropping process, performs a second cropping process on a portion of the object in a next frame, extracts a second edge degree from a second object image of the first object obtained through the second cropping process, compares the first edge degree and the second edge degree, and if the first edge degree is stronger than the second edge degree, detects object characteristic information based on the first object image, and if the second edge degree is stronger than the first edge degree, detects object characteristic information based on the second object image.

[0041] The edge degree of the first edge in the first object image is the first edge degree, and the edge degree of the second edge in the second object image is the second edge degree.

[0042] For example, the first edge degree and the second edge degree each represent how sharply (strongly) the edge appears, and can be expressed as one of a plurality of levels. For example, the first edge degree and the second edge degree each can be numerical values ​​representing how sharply (strongly) the edge appears.

[0043] A first edge degree being stronger than a second edge degree may mean that the first edge seen in the first object image is sharper than the second edge seen in the second object image.

[0044] A second edge degree being stronger than a first edge degree may mean that the second edge seen in the second object image is sharper than the first edge seen in the first object image.

[0045] In the traffic management system (100) according to embodiments of the present disclosure, the object may be a moving mobile object.

[0046] In the traffic management system (100) according to embodiments of the present disclosure, the image acquisition device (110) and the object detection device (120) can communicate in a wired or wireless manner.

[0047] In the traffic management system (100) according to embodiments of the present disclosure, the object is a vehicle, and the object characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0048] In a traffic management system (100) according to embodiments of the present disclosure, an object is a pedestrian, and object characteristic information may include information on one or more of the pedestrian's age, gender, and clothing.

[0049] The traffic management system (100) according to embodiments of the present disclosure may further include a traffic management device (130) that performs traffic management on the corresponding road by using object characteristic information detected by the object detection device (120).

[0050] For example, the object characteristic information may be vehicle characteristic information including vehicle license plate information. In this case, the traffic management device (130) may use the vehicle characteristic information including vehicle license plate information to identify a speeding vehicle or a vehicle that has violated a traffic signal to generate vehicle enforcement information. Alternatively, the traffic management device (130) may use the vehicle characteristic information including vehicle license plate information to determine whether the vehicle is a vehicle subject to enforcement registered with the National Police Agency Control Center, and if so, transmit the corresponding information to the National Police Agency Control Center.

[0051] For example, the object characteristic information may be pedestrian characteristic information including at least one of the pedestrian's age, gender, walking condition, clothing, and items worn. In this case, the traffic management device (130) may use the pedestrian characteristic information to control the signal system of the road or notify at least one of the pedestrian and the driver of a dangerous situation for the pedestrian. For example, if the traffic management device (130) determines that the pedestrian's walking condition is worse than a reference condition or slower than a reference speed through the pedestrian characteristic information, the traffic management device (130) may extend the display time of the traffic light allowing the pedestrian to cross the crosswalk beyond the reference time.

[0052] Referring to FIG. 2, in a traffic management system (100) according to embodiments of the present disclosure, an image acquisition device (110) may include at least one camera and may be installed on a road structure (200).

[0053] Referring to FIG. 2, in the traffic management system (100) according to embodiments of the present disclosure, the object detection device (120) may be located on the road structure (200) where the image acquisition device (110) is installed or at a location away from the road structure (200).

[0054] The image acquisition device (110) and the object detection device (120) may be separate devices or may be integrated into one device.

[0055] Referring to FIG. 2, in the traffic management system (100) according to embodiments of the present disclosure, the object detection device (120) and the traffic management device (130) may be separate devices or may be integrated into one device.

[0056] FIG. 3 shows an image for object characteristic detection according to embodiments of the present disclosure, and FIG. 4 shows a plurality of frames for object characteristic detection according to embodiments of the present disclosure.

[0057] Referring to FIGS. 3 and 4, the image (300) of the road acquired by the image acquisition device (110) is an image taken of the road over a certain period of time and may include a plurality of frames (F0, F1, F2, F3).

[0058] Referring to FIGS. 3 and 4, the object detection device (120) can extract the edge of the object (OBJ) from each of the plurality of frames (F0, F1, F2, F3). The object detection device (120) can compare the edges extracted from each of the plurality of frames (F0, F1, F2, F3) and select the frame (F1) in which the edge (410) of the object (OBJ) is expressed most strongly and clearly among the plurality of frames (F0, F1, F2, F3).

[0059] To be more specific, referring to FIGS. 3 and 4, each of the plurality of frames (F0, F1, F2, F3) may include an object image (IMG_OBJ). The object detection device (120) may crop a portion corresponding to the object (OBJ) in each of the plurality of frames (F0, F1, F2, F3) as an object image (IMG_OBJ), which is an image of the object (OBJ).

[0060] Referring to FIGS. 3 and 4, the object detection device (120) can extract the edge of the object (OBJ) from each of the plurality of object images (IMG_OBJ). The object detection device (120) can compare the edges extracted from each of the plurality of object images (IMG_OBJ) and select an object image (420) in which the edge (410) of the object (OBJ) is expressed most strongly and clearly among the plurality of object images (IMG_OBJ).

[0061] FIG. 5 is a block diagram of an object detection device (120) according to embodiments of the present disclosure.

[0062] Referring to FIG. 5, an object detection device (120) according to embodiments of the present disclosure may include an image storage unit (510) that stores an image of a road, an object detection unit (520) that detects an object through a machine learning algorithm based on the image, a frame selection unit (530) that selects a first frame from among a plurality of frames included in the image, and an object characteristic detection unit (540) that extracts an edge of an object from the first frame and detects object characteristic information about the object based on the extracted edge.

[0063] An object for object characteristic detection according to embodiments of the present disclosure may be a moving object.

[0064] Referring to FIG. 5, the object detection device (120) according to embodiments of the present disclosure may further include a communication unit (550) that communicates with the image acquisition device (110) in a wired or wireless manner and receives an image acquired from the image acquisition device (110).

[0065] The object detection unit (520) can detect an object including at least one of a vehicle and a pedestrian by executing deep learning as a machine learning algorithm based on an image.

[0066] The frame selection unit (530) can select a first frame having the highest sharpness information among the plurality of frames based on sharpness information for the edge of an object extracted from each of the plurality of frames.

[0067] Each of the multiple frames may be an object image that has cropped the portion of the object displayed in that frame.

[0068] The object characteristic detection unit (540) can extract the edge of an object from the first frame by executing an edge extraction algorithm or a machine learning algorithm, and detect object characteristic information based on the extracted edge.

[0069] For example, the object characteristic detection unit (540) can detect object characteristic information including vehicle license plate information when the object is a vehicle.

[0070] For another example, if the object is a vehicle, the object characteristic detection unit (540) can detect object characteristic information including at least one of vehicle size information and vehicle type information of the vehicle.

[0071] As another example, the object characteristic detection unit (540) can detect object characteristic information including information on one or more of the pedestrian's age, gender, walking status, clothing, and items worn when the object is a pedestrian.

[0072] FIG. 6 is a flowchart of an object detection method according to embodiments of the present disclosure.

[0073] Referring to FIG. 6, an object detection method according to embodiments of the present disclosure may include an image acquisition step (S610) of acquiring an image of a road through an image acquisition device (110), an object detection step (S620) of detecting an object through a machine learning algorithm based on the image by an object detection device (120), a frame selection step (S630) of selecting a first frame from among a plurality of frames included in the image by the object detection device (120), and an object characteristic detection step (S640) of extracting an edge of an object from the first frame by the object detection device (120) and detecting object characteristic information about the object based on the extracted edge.

[0074] In the object detection step (S620), the object detection device (120) can detect an object including at least one of a vehicle, a pedestrian, and a road structure by executing deep learning as a machine learning algorithm based on an image.

[0075] In the frame selection step (S630), the object detection device (120) can select a first frame having the highest sharpness information among the plurality of frames based on sharpness information for the edge of the object extracted from each of the plurality of frames.

[0076] Each of the multiple frames may be an object image that has cropped the portion of the object displayed in that frame.

[0077] An object for object characteristic detection according to embodiments of the present disclosure may be a moving object.

[0078] In an object detection method according to embodiments of the present disclosure, the object is a vehicle, and the object characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0079] In an object detection method according to embodiments of the present disclosure, the object is a pedestrian, and the object characteristic information may include information on one or more of the pedestrian's age, gender, walking status, clothing, and worn items.

[0080] An object detection device according to embodiments of the present disclosure may include an image storage unit that stores an image of a road; an object detection unit that detects an object through a machine learning algorithm based on the image; a frame selection unit that selects a first frame from among a plurality of frames included in the image; and an object characteristic detection unit that extracts an edge of an object from the first frame and detects object characteristic information about the object based on the extracted edge.

[0081] In the object detection device according to embodiments of the present disclosure, the object may be a moving object.

[0082] The object detection device according to embodiments of the present disclosure may further include a communication unit that communicates with an image acquisition device in a wired or wireless manner and receives an image acquired from the image acquisition device.

[0083] The object detection unit can detect an object including at least one of a vehicle and a pedestrian by executing deep learning as a machine learning algorithm based on an image.

[0084] The frame selection unit may select a first frame having the highest sharpness information among the plurality of frames based on sharpness information about the edges of the object extracted from each of the plurality of frames. Here, each of the plurality of frames may be an object image obtained by cropping a portion of the frame in which the object is displayed.

[0085] The object characteristic detection unit can extract the edge of the object from the first frame by executing an edge extraction algorithm or a machine learning algorithm, and detect object characteristic information based on the extracted edge.

[0086] The object characteristic detection unit can detect object characteristic information including vehicle license plate information when the object is a vehicle.

[0087] The object characteristic detection unit can detect object characteristic information including at least one of vehicle size information and vehicle type information when the object is a vehicle.

[0088] The object characteristic detection unit can detect object characteristic information including information on one or more of the pedestrian's age, gender, walking status, clothing, and items worn, if the object is a pedestrian.

[0089] An image-based object detection method according to embodiments of the present disclosure may include an image acquisition step of acquiring an image of a road through an image acquisition device; an object detection step of detecting an object through a machine learning algorithm based on the image by an object detection device; a frame selection step of selecting a first frame from among a plurality of frames included in the image by the object detection device; and an object characteristic detection step of extracting an edge of an object from the first frame by the object detection device and detecting object characteristic information about the object based on the extracted edge.

[0090] In the object detection step, the object detection device can detect an object including at least one of a vehicle, a pedestrian, and a road structure by executing deep learning as a machine learning algorithm based on an image.

[0091] In the frame selection step, the object detection device may select a first frame having the highest sharpness information among the plurality of frames based on sharpness information about the edge of the object extracted from each of the plurality of frames. Here, each of the plurality of frames may be an object image obtained by cropping a portion of the frame in which the object is displayed.

[0092] An object can be a moving object.

[0093] The object is a vehicle, the object characteristic information includes vehicle characteristic information of the vehicle, and the vehicle characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0094] The object is a pedestrian, the object characteristic information includes pedestrian characteristic information of the pedestrian, and the pedestrian characteristic information may include information about one or more of the pedestrian's age, gender, walking status, clothing, and items worn.

[0095] A traffic management system according to embodiments of the present disclosure may include an image acquisition device that acquires an image of a road; and an object detection device that detects object characteristic information about an object detected from the image based on the image.

[0096] The object detection device can detect objects in an image through deep learning.

[0097] The object detection device can perform a first cropping process on a portion of an object in a current frame, and extract a first edge degree from a first object image for the first object obtained according to the first cropping process.

[0098] The object detection device can perform a second cropping process on a portion of the object in the next frame, and extract a second edge degree from a second object image of the first object obtained according to the second cropping process.

[0099] The object detection device can compare a first edge degree and a second edge degree, and if the first edge degree is stronger than the second edge degree, detect object characteristic information based on the first object image, and if the second edge degree is stronger than the first edge degree, detect object characteristic information based on the second object image.

[0100] An object can be a moving object.

[0101] The image acquisition device and the object detection device can communicate in a wired or wireless manner.

[0102] The object is a vehicle, and the object characteristic information may include at least one of vehicle license plate information, vehicle size information, and vehicle type information.

[0103] The object is a pedestrian, and the object characteristic information may include information about one or more of the pedestrian's age, gender, walking status, clothing, and items worn.

[0104] According to the embodiments of the present disclosure described above, an image-based object detection method, an object detection device, and a traffic management system capable of detecting objects more accurately can be provided.

[0105] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of more accurately detecting objects using a machine learning algorithm such as deep learning can be provided.

[0106] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting object characteristic information representing characteristics of an object or characteristics of some components included in an object can be provided.

[0107] According to embodiments of the present disclosure, an image-based object detection method, an object detection device, and a traffic management system capable of accurately detecting a moving object and accurately detecting object characteristic information about the moving object can be provided.

[0108] The above description is merely an illustrative illustration of the technical concept of the present disclosure, and those skilled in the art will appreciate that various modifications and variations may be made without departing from the essential characteristics of the present disclosure. Furthermore, the embodiments disclosed in this disclosure are intended to illustrate, rather than limit, the technical concept of the present disclosure, and therefore, the scope of the technical concept of the present disclosure is not limited by these embodiments.

[0109]

[0110] CROSS-REFERENCE TO RELATED APPLICATION

[0111] This patent application claims priority under 35 USC §119(a) to Korean Patent Application No. 10-2023-0126202, filed September 21, 2023, the entire contents of which are incorporated herein by reference. Furthermore, this patent application claims priority in countries other than the United States for the same reasons, the entire contents of which are incorporated herein by reference.

Claims

1. Image storage unit that stores images of the road; An object detection unit that detects an object using a machine learning algorithm based on the above image; A frame selection unit for selecting a first frame from among a plurality of frames included in the above video; and An object detection device including an object characteristic detection unit that extracts an edge of the object from the first frame and detects object characteristic information about the object based on the extracted edge.

2. In paragraph 1, The above object is an object detection device that is a moving object.

3. In paragraph 1, An object detection device further comprising a communication unit that communicates with an image acquisition device in a wired or wireless manner and receives the image acquired from the image acquisition device.

4. In paragraph 1, The object detection unit is an object detection device that detects an object including at least one of a vehicle and a pedestrian by executing deep learning as the machine learning algorithm based on the image.

5. In paragraph 1, The frame selection unit selects the first frame having the highest sharpness information among the plurality of frames based on sharpness information for the edge of the object extracted from each of the plurality of frames, An object detection device in which each of the above plurality of frames is an object image obtained by cropping a portion of the frame in which the object is displayed.

6. In paragraph 1, An object detection device in which the object characteristic detection unit extracts the edge of the object from the first frame by executing an edge extraction algorithm or the machine learning algorithm, and detects the object characteristic information based on the extracted edge.

7. In paragraph 6, The object characteristic detection unit is an object detection device that detects object characteristic information including license plate information of the vehicle when the object is a vehicle.

8. In paragraph 6, The object characteristic detection unit is an object detection device that detects object characteristic information including at least one of vehicle size information and vehicle type information of the vehicle when the object is a vehicle.

9. In paragraph 6, The object characteristic detection unit is an object detection device that detects object characteristic information including information on at least one of the pedestrian's age, gender, walking status, clothing, and items worn when the object is a pedestrian.

10. Image acquisition step of acquiring images of a road using an image acquisition device; An object detection step in which an object detection device detects an object using a machine learning algorithm based on the image; A frame selection step in which the object detection device selects a first frame from among a plurality of frames included in the image; and An image-based object detection method, comprising an object characteristic detection step in which the object detection device extracts an edge of the object from the first frame and detects object characteristic information about the object based on the extracted edge.

11. In paragraph 10, An image-based object detection method in which, in the object detection step, the object detection device executes deep learning as the machine learning algorithm based on the image to detect the object including at least one of a vehicle, a pedestrian, and a road structure.

12. In paragraph 10, In the frame selection step, the object detection device selects the first frame having the highest sharpness information among the plurality of frames based on sharpness information about the edge of the object extracted from each of the plurality of frames, An image-based object detection method in which each of the above plurality of frames is an object image in which the portion in which the object is displayed in the frame is cropped.

13. In paragraph 10, The above object is an image-based object detection method that is a moving object.

14. In paragraph 10, An image-based object detection method wherein the object is a vehicle, the object characteristic information includes vehicle characteristic information of the vehicle, and the vehicle characteristic information includes at least one of license plate information, vehicle size information, and vehicle type information of the vehicle.

15. In paragraph 10, An image-based object detection method, wherein the object is a pedestrian, the object characteristic information includes pedestrian characteristic information of the pedestrian, and the pedestrian characteristic information includes information on at least one of the pedestrian's age, gender, walking status, clothing, and items worn.

16. An image acquisition device for acquiring images of a road; and Based on the above image, an object detection device is included that detects object characteristic information about an object detected from the image, The above object detection device, Detect the object in the above image through deep learning, Perform the first cropping process on the part of the object in the current frame, Extracting a first edge degree from a first object image of the first object obtained by the first cropping process, In the next frame, perform a second cropping process on the part of the above object, Extracting a second edge degree from a second object image of the first object obtained through the second cropping process, Compare the above first edge degree and the above second edge degree, If the first edge degree is stronger than the second edge degree, the object characteristic information is detected based on the first object image, A traffic management system that detects object characteristic information based on the second object image when the second edge degree is stronger than the first edge degree.

17. In paragraph 16, The above object is a traffic management system that is a moving object.

18. In paragraph 16, A traffic management system in which the image acquisition device and the object detection device communicate in a wired or wireless manner.

19. In paragraph 16, A traffic management system in which the object is a vehicle, and the object characteristic information includes at least one of vehicle license plate information, vehicle size information, and vehicle type information.

20. In paragraph 16, A traffic management system in which the object is a pedestrian, and the object characteristic information includes information on at least one of the pedestrian's age, gender, walking status, clothing, and items worn.

Images