KR102445138B1 - Image based speed control system using deep learning - Google Patents

Abstract

An image-based speed control system and method are disclosed. The speed control system according to an embodiment includes: acquiring image data including vehicle characteristic information using one camera or a plurality of cameras; learning a deep learning model using vehicle characteristic information included in the obtained image data; inputting image data for speed detection into the learned deep learning model; recognizing location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model; and detecting vehicle speed information by tracking location information related to the recognized vehicle characteristic information.

Description

Image-based speed control system using deep learning {IMAGE BASED SPEED CONTROL SYSTEM USING DEEP LEARNING}

The description below relates to speed detection technology.

With the entry into modern society, the amount of vehicles has increased rapidly, and at the same time, traffic congestion has also steadily increased. Various operation and control methods have been proposed accordingly. Recently, a vehicle is detected using various detection means and then managed and monitored using the detection data, which is called an Intelligent Transportation System (ITS).

Methods for detecting a vehicle include a method using a radar, a method using a camera, or a method using a loop detector. The method using radar is a method of identifying an object by analyzing the signal reflected when the signal transmitted from the radar collides with an object within the detection area of the radar. The method using radar is easy to grasp the relative position and speed of the vehicle and is less affected by the external environment. However, there is a big difference in performance depending on the type of radar and there is a problem of ghost generation according to the rotation of the signal.

The method using the camera is a method of analyzing and judging object information in the image information by using the image information to distinguish it from the background. The method using a camera is relatively inexpensive and easy to install compared to radar, but it is greatly affected by the external environment.

A deep learning model is trained using vehicle characteristic information included in image data obtained through at least one camera or a plurality of cameras, and vehicle speed information is detected from the image data for speed detection input to the learned deep learning model. systems and methods can be provided.

The speed detection method performed by the speed control system includes: acquiring image data including vehicle characteristic information using a single camera or a plurality of cameras; learning a deep learning model using vehicle characteristic information included in the obtained image data; inputting image data for speed detection into the learned deep learning model; recognizing location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model; and detecting vehicle speed information by tracking location information related to the recognized vehicle characteristic information.

In the learning step, from the image data, license plate information including a vehicle number or text configured on a license plate of a vehicle, a part of the license plate information, mark information of a manufacturer related to the vehicle, or a vehicle including an emblem of the vehicle It may include the step of training the deep learning model using the feature information.

The step of detecting the vehicle speed information may include calculating the lane information of the road displayed in the image data and correcting the distance information the vehicle moves through tracking the location information related to the recognized vehicle speed characteristic information. can

The inputting may include correcting camera shake by using road facility information or road facility information displayed in image data.

The one camera is that a first camera is installed at a preset position, and the detecting of the vehicle speed information includes setting an area based on a photographing range using the first camera installed at the preset position, and From the time the vehicle characteristic information enters the set area through the first camera until it leaves the set area, using frame information constituting image data including vehicle characteristic information acquired as the set area is photographed, the It may include detecting the recognized vehicle speed information.

The detecting of the vehicle speed information may include: based on a frame unit set in the first camera, vehicle characteristics according to time information from when the recognized vehicle characteristic information enters the set area to leave the set area The method may include detecting vehicle speed information through tracking of position information related to the recognized vehicle characteristic information using frame number information constituting the image data including the information.

The detecting of the vehicle speed information may include assigning an ID to a vehicle entering the set area through tracking of location information related to the recognized vehicle characteristic information.

The plurality of cameras include a first camera and a second camera installed vertically, the first camera is installed on the upper side, the second camera is installed on the lower side, and the step of detecting the vehicle speed information includes: The method may include detecting vehicle speed information based on distance information from the vehicle generated according to a height difference between the vertically installed first camera and the second camera.

In the detecting of the vehicle speed information, distance information calculated from the first camera to location information related to the recognized vehicle characteristic information and a distance calculated from the second camera to location information related to the recognized vehicle characteristic information may include: It may include the step of detecting the vehicle speed information in consideration of the error due to the installation angle of each camera and the error due to the processing speed of each camera in the information.

The detecting of the vehicle speed information may include determining violation information including lane violation information and speed violation information of the vehicle through location information related to the recognized vehicle characteristic information.

The obtaining of the image data may include obtaining image data by combining photographing data obtained by photographing using a camera and radar data obtained by using a radar.

The speed control system includes: an acquisition unit for acquiring image data including vehicle characteristic information using one camera or a plurality of cameras; a learning unit for learning a deep learning model using vehicle characteristic information included in the obtained image data; an input unit for inputting image data for speed detection into the learned deep learning model; a recognition unit for recognizing location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model; and a detection unit configured to detect vehicle speed information by tracking location information related to the recognized vehicle characteristic information.

A vehicle recognized from image data for speed detection input by learning a deep learning model using vehicle feature information included in image data acquired through at least one or more cameras or a plurality of cameras, and using the learned deep learning model By detecting the speed information of the vehicle, it is possible to reduce the error of vehicle recognition and detect the speed more accurately.

1 is a block diagram illustrating a configuration of a speed control system according to an embodiment.
2 is a flowchart illustrating a method of detecting vehicle speed information in a speed control system according to an exemplary embodiment.
3 and 4 are diagrams for explaining a method of detecting vehicle speed information using a plurality of cameras in a speed control system according to an embodiment.
5, 6, and 9 are diagrams for explaining a method of detecting the speed of a vehicle using a single camera in the speed control system according to an embodiment.
7 is a view for explaining the recognition of vehicle information in the speed control system according to an embodiment.
8 is an example for explaining how to perform Hangul misrecognition verification in the speed control system according to an embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

1 is a block diagram illustrating a configuration of a speed control system according to an exemplary embodiment, and FIG. 2 is a flowchart illustrating a vehicle speed detection method in the speed control system according to an exemplary embodiment.

The processor included in the speed control system 100 may include an acquisition unit 110 , a learning unit 120 , an input unit 130 , a recognition unit 140 , and a detection unit 150 . Such a processor and components of the processor may control the speed control system to perform steps 210 to 250 included in the vehicle speed detecting method of FIG. 2 . In this case, the processor and the components of the processor may be implemented to execute instructions according to the code of the operating system included in the memory and the code of at least one program. Here, the components of the processor may be expressions of different functions performed by the processor according to a control command provided by the program code stored in the speed regulation system 100 .

The processor may load the program code stored in the file of the program for the vehicle speed detection method into the memory. For example, when a program is executed in the speed control system 100 , the processor may control the speed control system to load a program code from a file of the program into the memory according to the control of the operating system.

In step 210 , the acquisition unit 110 may acquire image data including vehicle characteristic information using one camera or a plurality of cameras. The acquisition unit 110 may acquire image data by combining photographing data obtained by photographing using a camera and radar data obtained by using a radar.

In step 220 , the learning unit 120 may train the deep learning model by using vehicle characteristic information included in the acquired image data. The learning unit 120 uses vehicle characteristic information including license plate information, a part of license plate information, the manufacturer's mark information related to the vehicle, or the emblem of the vehicle from the image data, including the vehicle number or character configured on the license plate of the vehicle. Thus, deep learning models can be trained.

In step 230 , the input unit 130 may input image data for speed detection into the learned deep learning model. The input unit 130 may correct camera shake by using road facility information or road facility information displayed on image data. For example, the input unit 130 may correct camera shake by comparing pre-stored or externally stored road facility information and road facility information displayed in image data. Alternatively, the input unit 130 may correct the camera shake by checking the road facility information displayed in the image data. Accordingly, the input unit 130 may correct a portion in which the road, which is flat, is incorrectly viewed as uphill or downhill.

In step 240 , the recognizer 140 may recognize location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model. For example, the recognition unit 140 may recognize the vehicle information by learning the image data for speed detection in the deep learning model. The recognition unit 140 configures a deep learning model for object detection, and inputs image data for speed detection into the configured deep learning model and trains it to recognize vehicle information as a learning result. The recognition unit 140 may recognize vehicle identification information including the location, appearance, vehicle model, or vehicle number of the vehicle existing in the acquired image data.

In step 250 , the detection unit 150 may detect vehicle speed information by tracking location information related to the recognized vehicle characteristic information. The detection unit 150 may correct the distance information on which the vehicle moves by tracking the position information related to the recognized vehicle speed characteristic information by calculating the lane information of the road displayed in the image data.

For example, the detection unit 150 sets an area based on a photographing range using a first camera installed at a preset position, and detects the area set from when the vehicle characteristic information recognized through the first camera enters the set area. Recognized vehicle speed information may be detected using frame information constituting image data including vehicle characteristic information obtained as the set area is photographed until it leaves. The detection unit 150 determines the number of frames constituting image data including vehicle characteristic information according to time information from when the vehicle characteristic information enters the set area to the time it leaves the set area based on the frame unit set in the first camera. Vehicle speed information may be detected by tracking location information related to vehicle characteristic information recognized using the information. The detection unit 150 may assign an ID to a vehicle entering a set area based on the recognized vehicle characteristic information.

As another example, the detection unit 150 may detect vehicle speed information based on distance information from the vehicle generated according to a height difference between the vertically installed first camera and the second camera. The detection unit 150 includes the distance information calculated from the first camera to the position information related to the vehicle characteristic information recognized by the first camera and the distance information calculated from the second camera to the calculated distance information related to the vehicle characteristic information recognized by the installation angle of each camera. The vehicle speed information may be detected in consideration of the error caused by the camera and the error caused by the processing speed of each camera.

Also, the detection unit 130 may determine violation information including lane violation information and speed violation information of the vehicle through location information related to the recognized vehicle characteristic information. For example, it is possible to determine whether the vehicle violates a lane or speed by comparing it with pre-stored lane reference information and speed reference information using location information related to vehicle characteristic information recognized through a deep learning model. have. The detection unit 150 may determine whether the vehicle or vehicle speed information violates school zone regulations by comparing location information related to vehicle characteristic information based on reference information set in the school zone.

In addition, as the detection unit 150 analyzes the input image data for speed detection, the vehicle distribution information may be transmitted to differently control the signal data for distributing the traffic volume in association with the local signal system related to the image data. For example, when congestion occurs in a specific section of an area where the video data is captured as the video data is analyzed, vehicle distribution information may be transmitted in conjunction with a signal system of the area, and the specific section of the area through the signal system The signal data may be manipulated to pass the vehicle based on the vehicle distribution information.

3 and 4 are diagrams for explaining a method of detecting the speed of a vehicle using a plurality of cameras in the speed control system according to an embodiment.

The speed control system may constitute one camera or a plurality of cameras. If it is composed of one camera, it may be installed at a preset position, and if it is composed of a plurality of cameras, the first camera and the second camera may be installed vertically, one on the upper side and one on the upper side. It may be located on the lower side. For example, it is assumed that the first camera is installed on the upper side and the second camera is installed on the lower side of the first camera.

In addition, any one of a radar, a laser, and a lidar may be further installed in the speed control system. Radar is an abbreviation of Radio Detecting And Ranging, and it can check distance and direction by receiving electromagnetic waves reflected from objects after emitting electromagnetic waves. For example, the radar may acquire vehicle information by propagating a signal to a vehicle existing in the detection area, calculating a time at which the propagated signal is reflected, and calculating the distance and angle to the vehicle. As another example, the laser may measure an accurate distance by emitting a laser toward an object and detecting the reflected laser back. In this case, a bidirectional laser may be used to obtain vehicle information. As another example, Lidar not only has a similar name to radar, but also has a similar basic principle. However, by firing a laser with a different wavelength from the radar, the rider can recognize the surroundings by using the time and intensity of the reflected laser, changes in frequency, etc., and obtain more specific information. LiDAR can recognize anywhere as the sensor rotates 360 degrees. For example, a radar may be installed in a speed control system, and a radar function may be included in the camera.

The speed control system may acquire image data including vehicle characteristic information by using one camera or a plurality of cameras. Alternatively, the speed control system may acquire image data by combining photographing data obtained by photographing using at least one camera and radar data obtained using radar. For example, the speed control system may continuously take pictures using a plurality of cameras (a first camera and a second camera) and a radar. The final image data may be obtained by synchronizing the captured image data in this way.

The speed enforcement system can construct a deep learning model for object detection. The speed control system can construct a deep learning model for learning vehicle characteristic information. The speed control system may train a deep learning model using vehicle characteristic information included in image data obtained from a camera. The speed control system uses the vehicle characteristic information including the license plate information including the license plate number or text configured on the license plate of the vehicle from the image data, a part of the license plate information, the mark information of the manufacturer related to the vehicle, or the vehicle's emblem. A learning model can be trained. The speed control system can input image data for speed detection into the trained deep learning model. The speed control system may recognize location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model. The speed control system may detect vehicle speed information by tracking location information related to the recognized vehicle characteristic information.

In addition, the speed control system may recognize vehicle identification information including a location, an appearance, a vehicle model, and a vehicle number of an object existing in the image data for speed detection. Furthermore, the speed enforcement system may provide vehicle information including enforcement information about the vehicle. In this case, the enforcement information may mean for determining whether or not a rule set for each road is observed. The speed control system may recognize vehicle information by tracking location information related to vehicle characteristic information in the detection area. The speed control system derives vehicle enforcement information including whether the vehicle is driving normally and whether the vehicle is in lane violation through tracking location information related to vehicle characteristic information, and the derived vehicle enforcement information, and the vehicle related to the vehicle It is possible to provide vehicle information including location lane information, a vehicle model and a vehicle number. For example, the speed control system may determine the location of the vehicle, the shape of the vehicle or the number of the vehicle, etc. through tracking of location information related to vehicle characteristic information using image information for speed detection, and radar, laser and lidar Through the information, it is possible to determine the speed of the vehicle, the location of the vehicle, or the number of vehicles entering each lane. In addition, the speed control system may recognize the license plate of the vehicle, and may further estimate the car model. The speed control system can detect the speed of the vehicle and estimate the amount of traffic. The speed control system may recognize vehicle information through tracking of location information related to vehicle characteristic information, and when the vehicle is in the exclusive bus lane, it may determine a violation of regulations in the exclusive bus lane. Alternatively, the speed control system may determine vehicle speed information in the school zone based on the school zone information.

The speed control system may detect speed information of the vehicle based on distance information from the vehicle that is generated according to a height difference between the first camera and the second camera. The speed control system detects the speed information of the recognized vehicle based on difference information between the calculated distance information from the vertically installed first camera to the recognized vehicle and the calculated distance information from the second camera to the recognized vehicle. can For example, the speed control system may include the distance information calculated from the first camera to the position information related to the recognized vehicle characteristic information and the calculated distance information from the second camera to the position information related to the vehicle characteristic information recognized by each camera. Vehicle speed information may be detected in consideration of an error due to the installation angle of the , and an error due to the processing speed of each camera.

은 수학식 1를 통하여 계산될 수 있다.Specifically, as shown in FIG. 3 , the distance between the camera and the vehicle may be calculated with respect to an object (vehicle) existing to the left of the first camera and the second camera.

can be calculated through Equation (1).

Equation 1:

은 이미지 픽셀의 개수를 2로 나눈 값을 의미한다. where P ₁ is the pixel coordinate of the left camera (bottom camera) image,

is a value obtained by dividing the number of image pixels by 2.

는 수학식 2를 통하여 계산될 수 있다.Likewise,

can be calculated through Equation (2).

Equation 2:

The expression of Equation 3 can be derived from FIG. 1 .

Equation 3:

In this case, if the sine rule is applied, an expression such as Equation 4 may be derived.

Equation 4:

D represents the distance between the left camera (lower camera) and the vehicle. If Equation 4 is summarized and expressed for D, it can be expressed as Equation 5.

Equation 5:

The speed control system may consider that an error occurs with respect to the camera installation angle. For example, it is assumed that the installation heights of the two cameras are 5m and 6m, the distance from the camera installation post to the vehicle is 30m, and the camera angle of view is 70°.

If the distance R = 30 m from the camera to the vehicle,

가 도출될 수 있다.

can be derived.

과

를 도출하여 수학식 5에 대입되어 계산함에 따라 결과가 도출될 수 있다. Therefore,

class

A result can be derived by deriving and substituting it in Equation 5 for calculation.

에 -1°의 오차가 발생한 경우, At this time,

If an error of -1° occurs in

에 1°의 오차가 발생한 경우

If an error of 1° occurs in

This causes an error in the calculation result.

In addition, the speed control system may consider errors caused by the camera processing speed. As shown in FIG. 4 , in the case of 100 km/h, the moving speed is 27.7 m/sec, the image processing speed is 30 frames/sec, and the vehicle speed error is Ve = (d2-92.6cm)/(t1-t2). In this case, the maximum error of one frame may be 92.6 cm. In addition, when the speed is 100km/h and d2=30m, t1-t2=30m/27.7m/s=1.08sec, and a 3.1% error may occur at the speed Ve=96.9km/h.

5, 6, and 9 are diagrams for explaining a method of calculating a vehicle speed using one camera in the speed control system according to an exemplary embodiment.

In an embodiment, the speed control system may constitute one camera. In this case, one camera may be installed at a specific location. In addition, the speed detection and control system may be further installed any one of a radar, a laser, and a lidar.

The speed control system may acquire image data including vehicle characteristic information by using one camera or a plurality of cameras. The speed control system may train a deep learning model by using vehicle characteristic information included in the acquired image data.

The speed control system can input image data for speed detection into the trained deep learning model. The speed control system may recognize location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model. The speed control system may detect vehicle speed information by tracking location information related to the recognized vehicle characteristic information. In an embodiment, the speed control system may perform vehicle recognition from a point in time when a vehicle is photographed in a preset area 540 by one camera.

The speed control system may set an area 540 for detecting the speed of the vehicle. The speed control system may set an area 540 based on a photographing range using one camera set at a preset position. In this case, the size of the region 540 may be set by the user, and the size may be changed according to the user's input. For example, in the speed control system, the set area is photographed from the point in time when the vehicle characteristic information recognized through one camera enters (520) the set area (540) to the point at which it leaves (510) the set area (540). The speed information of the recognized vehicle 530 may be detected by using frame information constituting image data including the acquired vehicle characteristic information. In this case, the speed control system may assign an ID to the vehicle entering the set area 540 through tracking of location information related to the recognized vehicle characteristic information. The speed control system may detect vehicle speed information by tracking the ID of the vehicle for each frame based on the ID assigned to the vehicle.

The speed control system is based on the time information from when the vehicle characteristic information recognized based on the frame unit set in one camera enters (520) the set area 540 until it leaves (510) the set area 540. Recognized vehicle speed information may be detected using information on the number of frames constituting image data including vehicle characteristic information. In other words, the number of frames captured per second may be set in the camera. Accordingly, the speed control system may detect the vehicle speed by identifying location information (eg, coordinate information) related to vehicle characteristic information per frame. Here, the position information related to the vehicle characteristic information may mean a characteristic position corresponding to a part of the license plate, not the entire license plate of the vehicle, an emblem attached to the vehicle, mark information attached to the vehicle, and the like. For example, there may be a conversion table for providing the speed of the vehicle based on location information related to vehicle characteristic information determined by analyzing the frame. The speed control system may calculate the position difference of the vehicle between the first frame and the second frame, and extract the speed information present in the conversion table corresponding to the position difference based on the position difference between the position information of the vehicle characteristic information between the frames. can do. The speed control system may detect vehicle speed information. As another example, as another example, the speed control system may display a virtual line of a preset range in the area 540 . In this case, the virtual line may mean a plurality of lines set in the area to determine the speed of the vehicle based on the frame unit set in the camera, and the line may be displayed in a preset speed unit. For example, three virtual lines may be displayed in the area, and as the three virtual lines are displayed, four detailed areas may be generated. At this time, assuming that a difference of 25 km/h appears for each region, when the vehicle exists in the first region in the first frame and the vehicle exists in the second region in the second frame, the vehicle speed increases by 25 km/h can be judged. As each frame is compared, the speed control system may determine the speed information of the vehicle according to the distance the vehicle has moved through the location information related to the vehicle characteristic information based on the virtual line displayed in the region 540 .

Furthermore, the speed control system may correct the distance information the vehicle moves through the location information related to the recognized vehicle speed characteristic information by calculating the lane information of the road displayed in the image data. For example, the speed control system may calculate road lane information (eg, road width, how many lanes) displayed in the image data for speed detection.

Referring to FIG. 6 , if the time at which the first image data is captured is 00:07:15.241 and the time at which the second image data is captured is 00:07:16.385, the time difference is 1.141 seconds. By using the time difference information, it is possible to detect vehicle location information using a specific location related to vehicle characteristic information according to each frame to detect speed information about the vehicle.

Referring to FIG. 9 , an area based on a photographing range may be set using the first camera installed at a preset position. Frame information constituting image data including vehicle characteristic information acquired as the set area is photographed from when the location information related to the vehicle characteristic information recognized through the first camera enters the set area until it leaves the set area. It is possible to detect the recognized vehicle speed information using For example, first image data including vehicle characteristic information captured in a set area and second image data including vehicle characteristic information may be acquired. Then, the movement distance obtained by subtracting the recognized vehicle-specific information in the second image data (for example, the position of the number 3) and the recognized vehicle-specific information in the first image data, and the Vehicle speed information may be calculated using time information obtained by subtracting the photographing time and the photographing time of the first image data. Here, vehicle speed information may be calculated by calculating the moving distance (position '3' of the t2 image - the position of number '3' of the t1 image)/(t2-t1).

7 is a view for explaining the recognition of vehicle information in the speed control system according to an embodiment.

An algorithm for speed detection can be implemented as follows.

The speed enforcement system may detect vehicle speed information based on the implemented algorithm. The speed control system can set the size of the area. For example, the size of the region may be set based on a numerical value input by the user. The speed control system may recognize vehicle location information for each image data. In this case, a reference point may be set, and location information of the vehicle may be grasped based on the set reference point. The speed control system may acquire the center coordinates of the license plate from the first image data and the center coordinates of the license plate from the second image data. The speed control system may derive vehicle speed information based on the location information of the vehicle in each image data and the license plate center coordinates in each image data. Also, for example, the speed control system may recognize location information related to vehicle characteristic information using image data including vehicle characteristic information. In this way, location information of the vehicle may be estimated through location information related to the recognized vehicle characteristic information. By detecting vehicle speed information using a specific point (location information) according to vehicle characteristic information/vehicle characteristic information, the speed control system can increase recognition accuracy than recognizing the overall shape of the vehicle, and reduce speed detection errors. can

In addition, the speed control system may correct information on the distance the vehicle moves through location information related to vehicle speed characteristic information recognized by calculating lane information of a road displayed in the image data.

8 is an example for explaining how to perform Hangul misrecognition verification in the speed control system according to an embodiment.

When the speed control system recognizes location information related to vehicle characteristic information, misrecognition of characters on the license plate of the vehicle may occur. The license plate of the vehicle may be composed of letters and numbers of preset digits. For example, when the license plate of the vehicle is 45 and 1172, the speed control system may misrecognize the call as o, o, etc. In addition, among the characters present on the license plate of the vehicle, it may be mistakenly recognized as ha -> ah, mu -> u, bu -> mu, ru -> mu, and the like. Accordingly, the speed control system may correct the camera shake by using the facility information on the road.

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA). , a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions, may be implemented using one or more general purpose or special purpose computers. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that may include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. may be embodied in The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible by those skilled in the art from the above description. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (12)

A speed detection method performed by a speed control system, comprising:
obtaining image data including vehicle characteristic information by using a plurality of cameras;
learning a deep learning model using vehicle characteristic information included in the obtained image data;
inputting image data for speed detection into the learned deep learning model;
recognizing location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model; and
Detecting vehicle speed information by tracking location information related to the recognized vehicle characteristic information
including,
The learning step is
From the image data, using license plate information including a vehicle number or character configured on a license plate of a vehicle, a part of the license plate information, mark information of a manufacturer related to the vehicle, or vehicle characteristic information including the emblem of the vehicle Steps to train a learning model
including,
The input step is
Correcting camera shake using road facility information or road facility information displayed in image data
including,
The plurality of cameras, including a first camera and a second camera installed vertically, the first camera is installed on the upper side, the second camera is installed below the first camera,
The step of detecting the vehicle speed information includes:
Vehicle speed information is detected based on distance information from the vehicle that is generated according to a height difference between the vertically installed first camera and the second camera, and the first camera calculates location information related to the recognized vehicle characteristic information. Vehicle speed information in consideration of the error due to the installation angle of each camera and the error due to the processing speed of each camera in the distance information calculated from the distance information and the position information related to the recognized vehicle characteristic information from the second camera step to detect
A speed detection method comprising a. delete According to claim 1,
The step of detecting the vehicle speed information includes:
Compensating the distance information the vehicle moves through tracking the location information related to the recognized vehicle speed characteristic information by calculating the lane information of the road displayed in the image data
A speed detection method comprising a. delete delete delete delete delete delete According to claim 1,
The step of detecting the vehicle speed information includes:
and determining violation information including lane violation information and speed violation information of the vehicle through location information related to the recognized vehicle speed characteristic information. According to claim 1,
The step of obtaining the image data includes:
A step of acquiring image data by combining the imaging data acquired by using the camera and the radar data acquired using the radar
A speed detection method comprising a. In the speed control system,
an acquisition unit configured to acquire image data including vehicle characteristic information using a plurality of cameras;
a learning unit for learning a deep learning model using vehicle characteristic information included in the obtained image data;
an input unit for inputting image data for speed detection into the learned deep learning model;
a recognition unit for recognizing location information related to vehicle characteristic information from the input image data for speed detection using the learned deep learning model; and
A detection unit for detecting vehicle speed information by tracking location information related to the recognized vehicle characteristic information
including,
The learning unit,
From the image data, using license plate information including a vehicle number or character configured on a license plate of a vehicle, a part of the license plate information, mark information of a manufacturer related to the vehicle, or vehicle characteristic information including the emblem of the vehicle comprising training a learning model;
The input unit,
Compensating for camera shake by using road facility information or road facility information displayed in image data,
The plurality of cameras, including a first camera and a second camera installed vertically, the first camera is installed on the upper side, the second camera is installed below the first camera,
The detection unit,
Vehicle speed information is detected based on distance information from the vehicle that is generated according to a height difference between the vertically installed first camera and the second camera, and the first camera calculates location information related to the recognized vehicle characteristic information. Vehicle speed information in consideration of the error due to the installation angle of each camera and the error due to the processing speed of each camera in the distance information calculated from the distance information and the position information related to the recognized vehicle characteristic information from the second camera to detect
speed control system.

Images