KR101026852B1 - Method and system of supervision for car stop or parking - Google Patents

Abstract

PURPOSE: A system and method for detecting illegal vehicle parking are provided to determine whether vehicle parking is illegal by the appearance of a vehicle, thereby consecutively detecting illegal vehicle parking. CONSTITUTION: A camera unit(110) photographs a part or all of a parking control zone to obtain an image frame. A local controller(120) stores the image frame and controls the camera unit. A photographing module(100) includes the local controller. A central controller(200) receives image data and image frame from the photographing module and outputs the image data and image frame on a display unit(210). The central controller controls the photographing module.

Description

Method and System Of Supervision For Car Stop Or Parking}

The present invention relates to a system and method for detecting a parking violation, and more particularly, to obtain an image of a parking violation vehicle by photographing a portion or the whole range of the parking enforcement area, wherein a part of the exterior of the parking violation vehicle is covered by an obstacle. The present invention relates to a parking violation detection system and a method for determining whether a parking violation is determined only by information on the remaining part of the appearance of the parking violation vehicle even if the image is not taken or is distorted by a light.

In general, vehicles tend to park on either side of the road (or on the shoulder). This parked vehicle obstructs the flow of other vehicles while driving and substantially narrows the road. In addition, since such parking violations may lead to traffic accidents, the road traffic law has banned parking stops by setting various parking prohibition signs.

However, due to the fact that the personnel of the police officer who enforces the crackdown cannot be placed in every parking control area, the traffic lanes are overflowing even though there are stop signs on the road shoulders and the like.

In order to overcome this situation, traffic police are conducting more rapid patrols, but the effect is temporary, and collisions and frictions with civil complaints often occur during crackdowns.

In order to solve such a conventional problem, a part of the parking stop is performed by installing a camera and then installing a coil under the road or shooting by setting a photographing point of the vehicle using a laser.

However, in order to read whether the vehicle is a vehicle for violation, the shape or color information of the exterior of the vehicle must be acquired, and the vehicle for violation of the vehicle and the camera in the process of photographing the vehicle for violation of the vehicle in order to obtain the form and color information of the vehicle. If another vehicle passes by or stops temporarily due to a signal wait, and is inevitably obscured, or the color of the exterior of the parking vehicle is severely distorted by a lighting or headlight at night, An impossible case occurred.

In this case, due to the difference between the previously photographed image information and the image information captured in a state in which the image is obscured or the color is distorted, there is a problem that the image information for the stopped vehicle violation accumulated up to now is updated or deleted.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an object thereof is to provide a parking violation detection system and method capable of accurately determining whether a parking violation is accurate based only on information on a part of an appearance of a parking violation vehicle.

According to an aspect of the present invention, there is provided a parking stop detection system, including: a camera unit for capturing a part or entire range of the stopping area and obtaining an image frame; and storing the image frame and controlling the operation of the camera unit. And a photographing stage including a local control unit which extracts and registers the detected surveillance target and the image data of the surveillance region in the image frame. And a central control unit which receives the image data and the image frame from the image capturing terminal, collects the image data and the image frame, outputs them to a predetermined screen display means, or stores them on a predetermined storage medium, and controls the operation of the image capturing terminal.

Herein, the local controller binarizes each pixel of the image data corresponding to the surveillance target region occupied by the surveillance target among the image frames, and converts each pixel into N × M Grams (N rows and M columns). Grouped by N × M bundle units), and the monitoring target of the current video frame and the monitoring target of the existing video frame are the same based on the number of the same N × M Grams in the image data of the surveillance target region. You can determine if it is being monitored.

In addition, the local control unit, each of the binarized pixel of the image data for each 3 × 3 Grams (9 batch units consisting of three rows and three columns), 4 × 4 Grams (16 consisting of four rows and four columns) Group by N × M Grams or 5 × 5 Grams (25 rows of 5 rows and 5 columns).

In addition, the local controller compares an image frame currently acquired by the camera unit with an existing image frame, detects a pixel having a difference in luminance or RGB value, and then detects the detected pixel when the number of the detected pixels is greater than or equal to a predetermined scale. A grouping algorithm can be applied to the target to be monitored.

The local controller may further include the variable pixel and the luminance or the like when there is another pixel having a difference in luminance or an RGB value among four pixels adjacent to each other vertically from the variable pixel in which the luminance or RGB value difference occurs. The monitoring target area can be determined by designating different pixels in which the difference in RGB values have occurred as a group.

The local controller may compare the image frame including the current monitoring target with the image frame including the existing monitoring target, and compare the position of the current monitoring target with the position of the existing monitoring target. It is possible to register the monitoring target as a parking vehicle.

The local controller may extend the predetermined time limit when the motion is detected within the predetermined time limit from the monitoring target, and monitor the motion when the motion is not detected within the predetermined time limit from the monitoring target. Subjects can be registered as a vehicle for violation.

In addition, a predetermined eigenvalue is given to the monitored object, and when two or more monitored objects exist at the same time, the respective monitored objects are distinguished by using the predetermined eigenvalues, and then priority is given in order of magnitude or speed of the monitored object. By setting it, it is possible to preferentially track and photograph the monitoring target having a higher priority.

The camera unit may include a first camera for capturing a part or the whole range of the parking control area as the low resolution image, and capturing a high resolution image in all directions using a pan, tilt, and zoom function. It may include a second camera for tracking the registration plate of the registered vehicle to obtain the image frame of the license plate.

In addition, the camera unit, having a pan, tilt and zoom functions, to take a part or the entire range of the parking enforcement area, and to track the photographing license plate of the vehicle registered as a parking violation vehicle to obtain the image frame of the license plate It may include at least one third camera.

On the other hand, the parking violation detection method according to the present invention for achieving the above object, if the current image frame is obtained by capturing a part or the entire range of the parking stop area using the camera unit, if there is no previously stored image frame Registering the current video frame as an existing video frame to be compared; Compare the current video frame with a pre-stored existing video frame to detect a surveillance target and a surveillance target region, and if there is a detected surveillance target, extract and register the video data of the detected surveillance target and monitor the surveillance data among the video frames. Binarizing each pixel of the image data corresponding to the target region to be monitored by the target; Each binarized pixel is grouped and divided into N × M Grams (N × M group consisting of N rows and M columns) and based on the number of identical N × M Grams in the image data of the monitored region. Judging whether the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target; When the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target, it is determined whether each monitoring target is at the same position by comparing the positions occupied by each monitoring target in the current video frame and the existing video frame. It may include;

Here, when the movement is detected within the predetermined time limit from the monitoring target, the predetermined time limit is extended, and when the movement is not detected within the predetermined time limit, the monitoring target is moved to a vehicle for violation of vehicle speed. Registering; may further include.

In addition, in the step (b), by comparing the current video frame with a pre-stored existing video frame to detect the monitoring target, if there is no detected monitoring target may be updated (S308).

In addition, if it is determined that the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target, and not the same monitoring target, the monitoring target of a certain time elapsed may be deleted.

According to the system and method for detecting a parking violation according to the present invention, even if a part of the exterior of the parking violation vehicle is not covered by an obstacle and is not photographed or is distorted by a light, the parking violation may be performed only by information on the rest of the appearance of the parking violation vehicle. You can determine whether or not.

In addition, since the obtained image frame and the image information of the stop vehicle (monitoring target) are not collected and stored in the central control unit at a time, they are stored in each local control unit for a certain period of time, so that the data of the central control unit and the local control unit are stored. Even if there is a problem in transmission or the central control unit is temporarily inoperable, the image information about the parking violation zone is stored in the local control unit, so there is no effect of a gap in the parking violation enforcement.

1 is a block diagram of a parking violation detection system according to a preferred embodiment of the present invention,
2 is an exemplary diagram for explaining a grouping algorithm according to a preferred embodiment of the present invention;
3 and 4 are exemplary diagrams of pixels for explaining a method of grouping the binarized monitoring targets by N × M Grams according to a preferred embodiment of the present invention;
Figure 5 is a flow chart showing a parking violation detection method according to a preferred embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Prior to this, terms or words used in the present specification and claims should not be construed as being limited to the common or dictionary meanings, and the inventors should properly explain the concept of terms in order to best explain their own invention. Based on the principle that it can be defined, it should be interpreted as meaning and concept corresponding to the technical idea of the present invention.

Therefore, the embodiments described in the specification and the drawings shown in the drawings are only the most preferred embodiment of the present invention and do not represent all of the technical idea of the present invention, various modifications that can be replaced at the time of the present application It should be understood that there may be equivalents and variations.

First, the configuration of a parking violation detection system according to a preferred embodiment of the present invention will be described.

As shown in FIG. 1, the parking violation detection system according to the present invention is installed in a parking violation zone to detect a parking violation vehicle, and includes a camera unit 110 and a local controller 120. It may be provided including a photographing stage 100 and the central control unit 200.

Here, the camera unit 110 obtains an image frame by photographing a part or the entire range of the stop zone violation.

The camera unit 110 is fixed to a predetermined position as shown in Figure 1, the first camera 111 for capturing a part or the entire range of the stop zone violation in a low resolution image, pan, tilt and A second camera 112 may be used to capture an image of a high resolution in all directions by using a zoom function, and to track and photograph a license plate of a vehicle registered as a parking violation vehicle to obtain an image frame of the license plate.

Here, the predetermined position is a position designated by the designer of the system, and it is preferable that the predetermined position is located at a high place so as to easily photograph a large place.

In addition, a low resolution image refers to an image having a resolution of 640 × 480 or less, and may be an image of a level generally used in conventional CCTV, and the first camera 111 has a wide range of stops or road violations. Therefore, a plurality of dogs can be used.

That is, when the first camera 111 is installed on a straight road, two cameras may be provided on one side of the road in opposite directions so as to photograph both directions of the road. In a cross road such as a crossroad, the center of the intersection Four cameras may be provided to photograph all directions. In addition, when the photographing area in one direction is wide, a plurality of cameras may be provided to photograph the same direction as in FIG. 1, but photograph different photographing areas.

The second camera 112 captures an image of high resolution in all directions by using a pan, tilt, and zoom function. In this case, the second camera 112 does not always photograph an image, but photographs an image when a control command is received from the local controller 120 according to an analysis result of the image photographed from the first camera 111.

Here, the high resolution image is a concept contrasted with the image photographed by the first camera 111, but preferably has a resolution of 640x480 or more so that the user can sufficiently decode the content of the image, but is not limited thereto.

In addition, since the second camera 112 is capable of omnidirectional shooting, unlike the first camera 111 for capturing an image at a fixed position, the second camera 112 can sufficiently acquire an image for a parking violation zone, even if a plurality of cameras are not provided. At the same time, when two or more vehicle violations occur, a plurality of cameras may be provided to independently record each vehicle violation.

That is, the number of the first camera 111 varies depending on the photographing range of the parking violation zone, and the second camera 112 is included in the number of parking violation vehicles that can occur at the same time and the time involved in the parking violation object. The number provided is changed in consideration of the phosphorus stop violation time (usually 5 minutes).

In addition, the first camera 111 and the second camera 112 may be provided at the position where the image is taken, a sound input and output means for recording or outputting the sound of the field, such as a microphone or a speaker may be provided.

On the other hand, the local controller 120 stores the image frame obtained from the camera unit 110, controls the operation of the camera unit 110, the image of the monitored object and the monitoring target area detected in the image frame. Extract and register data.

The local controller 120 recognizes the monitoring target through the image photographed using the first camera 111 and registers it as a parking violation vehicle when the monitoring target is located at the same position for the parking violation time. Control the camera to track and photograph the monitoring target registered as the parking stop vehicle through the second camera 112, and capture the first camera 111 and the second camera 112. Save all video data shot through

In the present invention, by storing the image data in the local control unit 120 as well as the central control unit 200, it is possible to ensure reliable storage of the image data even if problems such as abnormalities of the communication network 130 in addition to the simple failure.

Here, the method of detecting the monitoring target from the image captured by the local controller 120 through the first camera 111 is as follows.

First, a pixel having a difference between two frames is detected by comparing the current image frame acquired by the first camera 111 with an existing image frame previously photographed, using luminance or RGB as an index, and then detecting the pixel. If the number of pixels is larger than a predetermined size, a monitoring target is detected by applying a grouping algorithm to the detected pixels.

Here, the predetermined scale is a configuration for filtering the fine noise data so that the second camera 112 does not track the object not intended by the user, such as fallen leaves.

The predetermined size is generally suitable for an area of about 1/2 of the average vehicle size, but the size of the object to be photographed also varies according to the position where the first camera 111 is fixedly installed, so that the predetermined size is set in consideration of this when the system is designed. It is desirable to be able to easily change it according to the user's setting.

In addition, when the monitoring target area occupied by the monitoring object in the image frame includes another pixel having a luminance or RGB difference among four pixels adjacent to each other vertically, vertically, or downward from a variable pixel that is a pixel in which the luminance or RGB value is different. By specifying the variable pixel and another pixel having a difference in luminance or RGB value as one group, the one group can be determined as one monitoring target, and a monitoring target region having an area equal to the space of the pixel can be determined. It can be determined.

For example, as shown in FIG. 2, (2,1), (2,3), and (3,2) are assigned to another pixel in which a difference in luminance or RGB value occurs among the pixels adjacent to the variable pixel (2,2). If it is determined based on (2,1) or (3,2) again, (3,1) becomes an adjacent pixel where a difference in luminance or RGB value occurs, and as a result, (2,1), (2,2) ), (2,3), (3,1) and (3,2) pixels constitute one group to be monitored, and the area to be monitored is (2,1), (2,2) Occupy as much space as (2,3), (3,1) and (3,2) pixels.

The grouping algorithm using the luminance or RGB value can detect the target and the target area, and the image data of the detected target and the target area is extracted and registered as the target in the local controller 120. .

In addition, the local controller 120 according to the present invention binarizes each pixel of the image data included in the registered monitoring target, and then each of the binarized pixel is monitored by the current video frame by N × M Grams and the existing video frame. It is determined whether the monitoring target of the same monitoring target is the same.

First, each pixel of image data corresponding to a surveillance target area occupied by a registered surveillance target in the image frame is binarized. That is, as shown in FIG. 2, (2,1), (2,2), (2,3), (3,1), and (3,2) pixels, which are to be monitored, are binarized.

Here, as a method of binarizing each pixel, an average binarization method of obtaining an average of all pixels in the image and setting the value as a threshold may be used, but the present invention is not limited thereto. Various kinds of binarization including in-P-tile (simple threshold method), mode, iterative binarization, adaptive binarization, otsu discretization binarization, triangular fuzzy binarization, rectangular fuzzy binarization, improved fuzzy binarization Manner may be applied.

However, in binarizing each pixel of the image data, if the threshold for binarization, i.e., the threshold value is set too low or too high, the discrimination power for distinguishing the monitoring target is lowered, so that the proper range of the threshold value is repeated through repeated experiments. It is preferable to select.

In addition, the local controller 120 divides each of the binarized pixels into groups by N × M Grams and classifies the current image frame based on the number of the same N × M Grams in the image data of the surveillance region. It is determined whether the target of the monitoring and the target of the existing video frame are the same.

Here, the N × M Grams is an element for presenting a criterion for determining whether each of the preset monitoring targets is the same in the system. The N × M Grams is a rectangular unit consisting of N rows and M columns. .

For example, for 3x3 Grams, a 9 pixel packed unit with 3 rows and 3 columns of squares, and 4x4 Grams with 16 rows and 4 columns of squares In the case of 5x5 grams, it means the unit of 25 pixels having the rectangular shape of 5 rows and 5 columns.

That is, each pixel of the image data occupied by the monitoring object detected by the local controller 120 is an area having 16 pixels in a 4 × 4 form as shown in FIG. 3, and is preset as shown in FIG. 3. When N × M Grams is 3 × 3 Grams and the monitoring target area is grouped into 16 pixels, the detected monitoring targets are (a), (b), (c), (d), etc. as shown in FIG. You will have four 3 × 3 Grams.

Accordingly, the monitoring target of the current video frame and the monitoring target of the existing video frame are based on the number of the same 3 × 3 grams present in the video data of each of the two video frames, the current video frame and the existing video frame. It can be determined whether it is the same monitoring target. That is, if there are more than a certain number of identical 3 × 3 Grams in two images (the monitoring target area of the current video frame and the monitoring target area of the existing video frame), the monitoring target of the current video frame and the monitoring target of the existing video frame are the same. Will be judged.

In the case of the 3 × 3 Grams, each of 16 pixels may form a 3 × 3 Gram of 512 (2 ^{3 × 3} ) types. Thus, when the size of the constants N and M is preset More various types of N × M Gram, but if the constant size of N and M is set too low, the discriminating power of the criterion is weakened, and if the constant size of N and M is set too high, the same monitoring target Overloading may occur. Therefore, it is preferable that the magnitudes of the constants of N and M are determined within an appropriate range in consideration of the discriminating power of the criterion and the processing capability of the system.

In this way, the local controller 120 based on a result of comparing the position of the surveillance target region occupied by the surveillance target among the current video frame with the position of the surveillance subject region occupied by the surveillance target among the existing video frames, If the location of the monitored area is the same for more than a predetermined time limit, the corresponding monitored object is registered as a parking vehicle.

In addition, when the movement is detected from the monitoring target within the predetermined time limit, the predetermined time limit is extended, and when the movement is not detected within the predetermined time limit from the monitoring target, the parking target is violated. Register as a vehicle.

Here, the predetermined time limit means a parking allowable time allowed in the parking violation zone. For example, if vehicle A stops at 17:00:00 at the parking violation zone, and moves from its current location before 17:04:59, it is not registered as a parking violation vehicle but is detected within the parking violation zone. The predetermined time limit is given again, and when the time exceeds 17:05:00 without moving from the current position, the local controller 120 determines the vehicle A as a parking violation vehicle.

In addition, when two or more monitored objects are detected at the same time, the local controller 120 assigns a predetermined eigenvalue to each monitored object, and distinguishes each monitored object by using the predetermined eigenvalue to determine the size of the monitored object. Alternatively, it is preferable to set the priority in the order of the moving speed, to preferentially track and photograph the monitoring target having the higher priority.

In this case, when the monitoring target is registered as a parking violation vehicle, the local controller 120 operates the second camera 112 to track and photograph the license plate of the vehicle registered as the parking violation vehicle to obtain an image frame of the license plate.

Meanwhile, the central controller 200 receives the image data and the image frames from the photographing terminal 100, collects them, outputs them to a predetermined screen display means 210, or stores them in a predetermined storage medium. Performs a function of controlling the overall operation of the photographing stage (100).

Next, a parking violation detection method using a parking violation detection system according to a preferred embodiment of the present invention will be described with reference to FIG. 5.

First, as shown in FIG. 5, the camera unit 110 acquires a current image frame by capturing a portion or the entire range of the parking stop area (S301), but if there is no existing image frame previously stored, The current video frame is registered as an existing video frame to be compared (S303).

Subsequently, when there is a pre-registered existing video frame, by comparing the current video frame and the pre-stored existing video frame is detected (S307), if there is a detected monitoring target, the detected Extracting and registering the image data of the surveillance target and binarizing each pixel of the image data corresponding to the surveillance region occupied by the surveillance target in the image frame (S309).

Here, the monitoring target is detected by comparing the current video frame with a previously stored existing video frame, and when there is no detected monitoring target, the current video frame is updated with a new video frame.

Subsequently, the binarized pixels of the image data are grouped by N × M Grams (N × M group consisting of N rows and M columns), and the same N × M Grams are included in the image data of the surveillance region. Comparing the number of the existing (S311) to determine whether the monitoring target of the current video frame and the monitoring target of the existing video frame is the same monitoring target (S313).

Here, it is preferable to determine whether the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target, and if the monitoring target of the current video frame is not the same monitoring target, delete the monitoring target after a predetermined time has elapsed.

Also, when the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target, each monitoring target is compared to a position occupied by each monitoring target in the current video frame and the existing video frame. It is determined whether there is (S317).

Subsequently, when the motion is detected within the predetermined time limit from the monitoring target, the predetermined time limit is extended. When the movement is not detected within the predetermined time limit, the monitoring target is moved to a vehicle for violation of vehicle speed. Register (S319).

Here, when the monitoring target is registered as a vehicle for violation, the local controller 120 operates the second camera 112 to track and photograph the license plate of the vehicle registered as the vehicle for violation. do.

As described above, although the present invention has been described by way of limited embodiments and drawings, the present invention is not limited thereto and is intended by those skilled in the art to which the present invention pertains. Of course, various modifications and variations are possible within the scope of equivalents of the claims to be described.

100 Cameras 110 Cameras
111 The first camera 112 The second camera
120 Local control unit 130 Communication network
200 ... central control unit 210 ... display means

Claims (12)

In a parking violation detection system that is installed in a parking enforcement area to detect a parking violation vehicle,
By using a first camera 111 and pan, tilt and zoom functions to capture a part or the entire range of the parking enforcement area as a low-resolution image, to shoot a high-resolution image in all directions, license plate of a vehicle registered as a parking violation A camera unit 110 including a second camera 112 for acquiring an image frame of the license plate and storing the image frame, controlling the operation of the camera unit 110, and detecting the image frame. A photographing stage (100) including a local controller (120) for extracting and registering image data of the monitored object and the monitored object region;
The image data and the image frame are received from the photographing terminal 100, collected and output to the predetermined screen display means 210 or stored in a predetermined storage medium, and the operation of the photographing terminal 100 is controlled. It includes; Central control unit 200,
The local controller 120 binarizes each pixel of the image data corresponding to the surveillance target area occupied by the surveillance target among the video frames.
Each binarized pixel of the image data is divided by 3 × 3 Grams (9 groups of 3 rows and 3 columns), 4 × 4 Grams (16 groups of 4 rows and 4 columns) or 5 × Grouped by one selected N × M Grams among 5 Grams (25 groups of 5 rows and 5 columns), and based on the number of the same N × M Grams present in the image data of the monitored area. Determine whether the monitoring target of the current video frame and the monitoring target of the existing video frame are the same monitoring target, or
After comparing the current image frame with the existing image frame through the camera unit 110 and detecting a pixel having a difference in luminance or RGB value, four adjacent pixels are arranged up, down, left, and right from a variable pixel that is a pixel having a difference in luminance or RGB value. If there is another pixel having a difference in luminance or RGB value among the pixels, the monitoring target area is determined by designating the variable pixel and another pixel in which the difference in luminance or RGB value occurs as one group,
On the basis of a result of comparing the position of the surveillance object area occupied by the surveillance object in the current video frame with the position of the surveillance object area occupied by the surveillance object in the existing video frame, the movement is reduced within a predetermined time limit from the surveillance object. If it is detected, the predetermined time limit is extended, and if the motion is not detected within the predetermined time limit from the monitoring target, the monitoring target is registered as a parking violation vehicle,
When two or more monitored objects are detected at the same time, a predetermined eigenvalue is given to each monitored object, and the priority is set in order of the size or moving speed of the monitored object by distinguishing each monitored object using the predetermined eigenvalues. The vehicle stopping violation detection system according to claim 1, wherein the monitoring target having a higher priority is tracked and photographed.






delete delete delete delete delete delete delete delete delete delete delete

Images