CN110047092B - multi-target real-time tracking method in complex environment - Google Patents

Abstract

The invention discloses a multi-target real-time tracking method in a complex environment, comprising a first video acquisition module, a second video acquisition module, an information transmission module, a database, a background clipping module, a video processing module, a display module, a controller and an alarm module ; In the process of cutting out the static background, the present invention collects a large number of video frame images of the monitoring area, and integrates a large number of video frame images to reduce the influence of character activities and accidental events on the background clipping, and simultaneously affect the monitoring area. Objects with a certain range of motion that appear in the system are analyzed to reduce the noise generated by the moving object during the tracking process, thereby reducing the difficulty of analysis and improving the accuracy of analysis and tracking; the present invention also establishes a coordinate system for the global monitoring area. The corresponding coordinate points are read, and the relay tracking is performed on the target according to the coordinates of the target, so as to prevent the target from being lost during the tracking process of the target, and the relay cannot be completed in time.

Description

A real-time multi-target tracking method in complex environment

technical field

The invention belongs to the technical field of video monitoring, and in particular relates to a multi-target real-time tracking method in a complex environment.

Background technique

With the progress of society and the development of cities, a large number of video surveillance systems are installed in both public and private areas. Video surveillance systems are an important means of mastering public security management. The establishment of video surveillance systems is of great significance to maintaining social stability. .

The video surveillance technology in the traditional technology mainly plays the role of collecting, browsing and recording. For the monitoring field, manual inspection and judgment are required, and manual inspection and judgment require the staff to be in a state of high concentration for a long time. And this is obviously more difficult, so it has also spawned a large number of smart monitoring systems. The smart monitoring system mainly tracks and manages the target after the target is calibrated. When the target moves abnormally during the tracking process, an alarm is issued. It can be seen that no matter what kind of smart monitoring system, the tracking and monitoring of the target is a necessary technology, but in the existing technology, the smart tracking and monitoring technology still has its shortcomings. In a complex environment, the monitoring camera is easily and frequently blocked Blocked by the environment such as buildings, the monitoring camera cannot directly track the target, and it is easy to follow the wrong or lost track of the monitoring target. Moreover, in the actual monitoring environment, the environment itself is changeable, such as fallen leaves, Foregrounds such as insects and birds will bring troubles to the extraction of the background, thereby affecting the extraction of the background and further affecting the tracking of the target. In order to solve this problem, the present invention provides the following technical solutions.

Contents of the invention

The purpose of the present invention is to provide a multi-target real-time tracking method in a complex environment.

The technical problem to be solved in the present invention is:

1. The important point of tracking and monitoring is to cut out the static background, and then subtract the background from the obtained image to get the foreground object, and then monitor it. However, in actual operation, there are many foreground objects that we do not need to monitor. , such as fallen leaves, insects, etc. When these foreground objects are covered or intersected with the target that needs to be monitored, it will affect the monitoring of the target, and when updating the background, when these foreground objects exist, it will also affect the clipping of the background. influences;

2. In the prior art, when the background environment is relatively complex, the monitoring target is often covered by buildings, etc., resulting in the inability to track and monitor the monitoring target in time, which is prone to the phenomenon of lost and wrong tracking.

The purpose of the present invention can be achieved through the following technical solutions:

A multi-target real-time tracking method in a complex environment, comprising a first video acquisition module, a second video acquisition module, an information transmission module, a database, a background clipping module, a video processing module, a display module, a controller and an alarm module;

Both the first video capture module and the second video capture module are surveillance cameras, which are used to collect video image information, and transmit the collected video image information to the database, background clipping module and video processing through the information transmission module module, wherein the monitoring area of the first video acquisition module is the same as that of the second video acquisition module and the monitoring areas of the first video acquisition module and the second video acquisition module are unchanged, the first video acquisition module monitors the overall situation of the monitoring area, and the second The second video acquisition module can track and monitor the monitoring targets in the monitoring area by changing the focal length;

The database is used to store the video image information transmitted by the video acquisition module;

The background cutting module is used to cut out and update the background of the monitoring screen of the first video acquisition module, and transmit the updated background to the video processing module;

The method that the background cutting out module cuts out and updates the background of the monitoring picture of the first video acquisition module is:

Step 1. Within a period of time, read the video frame images of the monitoring video information uploaded by the first video acquisition module every preset time T1, thereby obtaining ²ⁿ video frame images, wherein n≥1;

Step 2. When comparing the two pictures, use the preset fixed position object in the video monitoring area as a reference to read the corner points of the object in each video frame, and use at least two of the object's The corner points establish a plane coordinate system;

Step 3. Compare every two video frame images, cut off the part that only appears in one video frame image, and superimpose the two video frame images, and continue to superimpose the superimposed images in pairs, and finally Obtain an image, and use the plane coordinate system established on different video frame images as the corresponding standard when comparing;

Step 4, a plurality of video frame images obtained in step 1 are divided into a plurality of regions with the coordinate origin as corresponding points, and the positions of the divided regions on each video frame image are the same, when in at most four adjacent regions There is the same object, and whether it is the same object is judged by the color and color distribution of the object, then the object is calibrated, and the read 2 ⁿ video frame images are superimposed to obtain the motion range of the object, and the The range of motion of the object is cut and added to the corresponding coordinate position in the image obtained in step 3 to obtain a background image, and the movement within the range of motion of the object is shielded during tracking analysis;

Step 5. Perform the operations from step 1 to step 4 with the preset time T2 as a cycle, and transmit the obtained background image to the video processing module to replace the background image of the previous cycle;

The video processing module is used to read the video image information collected by the first video acquisition module and the background image uploaded by the background cutting module, and analyze and track the targets within the monitoring range according to the video image information and the background image. When the video processing module analyzes and obtains that the monitoring target has a preset feature, the controller controls the alarm module to send an alarm message, and the preset feature refers to the feature data of the tracking target preset in the database;

The display module is used to display the monitoring video information of the monitoring area processed by the video processing module.

As a further solution of the present invention, the video processing module analyzes the video images collected by the first video acquisition module, and obtains the traffic information of the area monitored by the first video acquisition module. , the background clipping module reads the video frame image of the surveillance video information uploaded by the first video acquisition module.

As a further solution of the present invention, the video processing module controls the working method of the second video acquisition module as follows:

S1. Read the video image information collected by the first video capture module and the background image uploaded by the background clipping module;

S2. Read the real-time coordinate position of the preset object in the background image. Through the coordinate position, the video processing module obtains the corresponding focal length value and transmits it to the second video acquisition module;

S3. The second video acquisition module adjusts its own focal length according to the received focal length value to obtain a clear tracking image.

As a further solution of the present invention, the method for the video processing module to track and monitor the target is:

SS1, with a first video acquisition module and a second video acquisition module as a group of video acquisition modules, multiple groups of video acquisition modules are set in the global monitoring area, and at least two groups of adjacent video acquisition modules are in the local monitoring area There are partial overlaps to ensure that the global monitoring area has no monitoring dead ends;

SS2. Establish a plane coordinate system with any point in the global monitoring area as the origin, and mark the overlapping area of at least two adjacent local monitoring areas;

SS3, a video acquisition module performs real-time continuous tracking and shooting of the tracked target according to its path, motion velocity vector and image recognition algorithm, and the video acquisition module calculates the real-time coordinate information of the tracked target in the monitoring area of the video acquisition module ;

SS4. When the tracked target enters the overlapping area, the video processing module simultaneously receives the video monitoring information of the local monitoring area involved in the overlapping area, and predicts its traveling direction in the overlapping area according to the speed vector of the tracked target, and the corresponding direction The video acquisition module is ready to carry out relay tracking. When the video acquisition module in the predicted direction fails to acquire the corresponding target outside the overlapping area, the original video acquisition module for target tracking continues to track.

Beneficial effects of the present invention:

1. In the process of cutting out the static background, the present invention collects a large number of video frame images of the monitoring area, and integrates a large number of video frame images to reduce the influence of human activities and accidental events such as fallen leaves on the background clipping, and at the same time Analyze the objects with a certain range of activities in the monitoring area, reduce the noise generated by the active objects during the tracking process, thereby reducing the difficulty of analysis and improving the accuracy of analysis and tracking;

2. The present invention establishes a coordinate system for the global monitoring area, reads the coordinate points corresponding to the position of the target, and performs relay tracking on the target according to the coordinates of the target, preventing the target from being lost during the tracking process of the target, and being unable to Complete the relay in time.

Description of drawings

The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.

Fig. 1 is a schematic diagram of the system structure of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

A multi-target real-time tracking method in a complex environment, as shown in Figure 1, comprising a first video acquisition module, a second video acquisition module, an information transmission module, a database, a background clipping module, a video processing module, a display module, and a controller and alarm module;

Both the first video capture module and the second video capture module are surveillance cameras, which are used to collect video image information, and transmit the collected video image information to the database, background clipping module and video processing through the information transmission module module, wherein the monitoring area of the first video acquisition module is the same as that of the second video acquisition module and the monitoring areas of the first video acquisition module and the second video acquisition module are unchanged, the first video acquisition module monitors the overall situation of the monitoring area, and the second The second video acquisition module can track and monitor the monitoring target in the monitoring area by changing the focal length, so as to ensure that a clear image of the monitoring target can be obtained;

The database is used to store the video image information transmitted by the video acquisition module, which is convenient for later reading;

The background cutting module is used to cut out and update the background of the monitoring screen of the first video acquisition module, and transmit the updated background to the video processing module;

The method that the background cutting out module cuts out and updates the background of the monitoring picture of the first video acquisition module is:

Step 1. Within a period of time, read the video frame images of the monitoring video information uploaded by the first video acquisition module every preset time T1, thereby obtaining ²ⁿ video frame images, wherein n≥1;

Step 2. When comparing the two pictures, use the preset fixed position object in the video monitoring area as a reference to read the corner points of the object in each video frame, and use at least two of the object's The corner points establish a plane coordinate system;

Step 3. Compare every two video frame images, cut off the part that only appears in one video frame image, and superimpose the two video frame images, and continue to superimpose the superimposed images in pairs, and finally Obtain an image, and use the plane coordinate system established on different video frame images as the corresponding standard when comparing;

Step 4, a plurality of video frame images obtained in step 1 are divided into a plurality of regions with the coordinate origin as corresponding points, and the positions of the divided regions on each video frame image are the same, when in at most four adjacent regions There is the same object, and whether it is the same object is judged by the color and color distribution of the object, then the object is calibrated, and the read 2 ⁿ video frame images are superimposed to obtain the motion range of the object, and the The range of motion of the object is cut and added to the corresponding coordinate position in the image obtained in step 3 to obtain a background image, and the movement within the range of motion of the object is shielded during tracking analysis;

Step 5: Perform the operations from step 1 to step 4 with the preset time T2 as a cycle, and transmit the obtained background image to the video processing module to replace the background image in the previous cycle.

This step can eliminate the interference to the background clipping of objects such as flags and branches that have been in the monitoring image for a long time but whose positions will change to a certain extent in the monitoring screen, because objects such as flags and branches will shake significantly , when the background clipping module reads a video frame image as the background, it cannot determine the shaking range of this type of object, so when performing information processing, this type of object will also be regarded as the target moving object by the video processing module. Processing, increased interference, increased processing difficulty.

As a further solution of the present invention, in order to improve the automation intensity of the real-time tracking system, the video processing module is also used to analyze the video images collected by the first video collection module, and obtain the flow of people in the monitored area of the first video collection module information, when the flow of people in the monitoring area was less than the preset value, the background clipping module read the video frame image of the monitoring video information uploaded by the first video acquisition module, thereby reducing the background clipping out of the video frame image read by the module Character images, reducing the difficulty of the background cutting process.

The video processing module is used to read the video image information collected by the first video acquisition module and the background image uploaded by the background cutting module, and analyze and track the targets within the monitoring range according to the video image information and the background image. When the video processing module analyzes and obtains that the monitoring target has a preset feature, the controller controls the alarm module to send an alarm message, and the preset feature refers to the feature data of the tracking target preset in the database;

The working method of the second video acquisition module controlled by the video processing module is:

S1. Read the video image information collected by the first video capture module and the background image uploaded by the background clipping module;

S2. The coordinate position in the background image has a one-to-one correspondence with the distance between the position and the video capture module, and the distance between the position and the video capture module has a corresponding relationship with the focal length of the camera. Read the preset object in the background The real-time coordinate position in the image, through the coordinate position, the video to be processed module obtains the corresponding focal length value and transmits it to the second video acquisition module;

S3. The second video acquisition module adjusts its own focal length according to the received focal length value, so as to obtain a clear tracking image.

The method for the video processing module to track and monitor the target is:

SS1, with a first video acquisition module and a second video acquisition module as a group of video acquisition modules, multiple groups of video acquisition modules are set in the global monitoring area, and at least two groups of adjacent video acquisition modules are in the local monitoring area There are partial overlaps to ensure that the global monitoring area has no monitoring dead ends;

SS2. Establish a plane coordinate system with any point in the global monitoring area as the origin, and mark the overlapping area of at least two adjacent local monitoring areas;

SS3, a video acquisition module performs real-time continuous tracking and shooting of the tracked target according to its path, motion velocity vector and image recognition algorithm, and the video acquisition module calculates the real-time coordinate information of the tracked target in the monitoring area of the video acquisition module ;

SS4. When the tracked target enters the overlapping area, the video processing module simultaneously receives the video monitoring information of the local monitoring area involved in the overlapping area, and predicts its traveling direction in the overlapping area according to the speed vector of the tracked target, and the corresponding direction The video acquisition module is ready to carry out relay tracking. When the video acquisition module in the predicted direction fails to acquire the corresponding target outside the overlapping area, the original video acquisition module for target tracking will continue to track to prevent the target from being lost.

The display module is used to display the monitoring video information of the monitoring area processed by the video processing module.

The above content is only an example and description of the structure of the present invention. Those skilled in the art make various modifications or supplements to the described specific embodiments or replace them in similar ways, as long as they do not deviate from the structure of the invention or Anything beyond the scope defined in the claims shall belong to the protection scope of the present invention.

Claims (4)

1. A multi-target real-time tracking method under a complex environment is characterized by comprising a first video acquisition module, a second video acquisition module, an information transmission module, a database, a background clipping module, a video processing module, a display module, a controller and an alarm module;

The first video acquisition module and the second video acquisition module are both monitoring cameras and are used for acquiring video image information and transmitting the acquired video image information to the database, the background clipping module and the video processing module through the information transmission module, wherein the monitoring areas of the first video acquisition module and the second video acquisition module are the same, the monitoring areas of the first video acquisition module and the second video acquisition module are unchanged, the first video acquisition module monitors the whole situation of the monitoring area, and the second video acquisition module can track and monitor a monitored target in the monitoring area by changing the focal length;

the database is used for storing the video image information transmitted by the video acquisition module;

The background clipping module is used for clipping and updating the background of the monitoring picture of the first video acquisition module and transmitting the updated background to the video processing module;

The method for the background cutting-out module to cut out and update the background of the monitoring picture of the first video acquisition module comprises the following steps:

step one, reading a video frame image of monitoring video information uploaded by a first video acquisition module at preset time intervals of T1 within a period of time, and obtaining a result 2ⁿa video frame image of a page, wherein n is more than or equal to 1;

when two pictures are compared, reading the angular points of the object in each video frame picture by taking the object with a preset fixed position in a video monitoring area as a reference, and establishing a plane coordinate system by using at least two angular points of the object;

comparing every two video frame images, cutting off the part which only appears in one video frame image, overlapping the two video frame images, continuously overlapping the two overlapped images in pairs to obtain an image, and taking a plane coordinate system established on different video frame images as a corresponding standard during comparison;

Step four, dividing the plurality of video frame images obtained in the step one into a plurality of areas by taking the origin of coordinates as corresponding points, wherein the positions of the divided areas on each video frame image are the same, when the same object exists in at most four adjacent areas and whether the same object is judged by the color and the color distribution of the object, calibrating the object, and reading the 2ⁿSuperposing the video frame images to obtain the motion range of the object, cutting the motion range of the object, adding the cut motion range of the object into the corresponding coordinate position in the image obtained in the third step to obtain a background image, and shielding the motion in the motion range of the object when tracking analysis is carried out;

Step five, performing the operations from the step one to the step four by taking the preset time T2 as a period, and transmitting the obtained background image to a video processing module to replace the background image in the previous period;

The video processing module is used for reading the video image information acquired by the first video acquisition module and the background image uploaded by the background clipping module, analyzing and tracking a target in a monitoring range according to the video image information and the background image, and when the video processing module analyzes that the monitored target has preset characteristics, the controller controls the alarm module to send alarm information, wherein the preset characteristics refer to characteristic data of a tracking target preset in the database;

the display module is used for displaying the monitoring video information of the monitoring area processed by the video processing module.

2. the method for multi-target real-time tracking in a complex environment according to claim 1, wherein the video processing module analyzes the video image acquired by the first video acquisition module to obtain the pedestrian flow information of the area monitored by the first video acquisition module, and when the pedestrian flow of the monitored area is smaller than a preset value, the background clipping module reads the video frame image of the monitored video information uploaded by the first video acquisition module.

3. the multi-target real-time tracking method under the complex environment according to claim 1, wherein the working method of the video processing module for controlling the second video acquisition module is as follows:

S1, reading the video image information collected by the first video collection module and the background image uploaded by the background cutting-out module;

s2, reading the real-time coordinate position of the preset object in the background image, and transmitting the corresponding focal length value obtained by the video to-be-processed module to the second video acquisition module through the coordinate position;

and S3, the second video acquisition module adjusts the focal length of the second video acquisition module according to the received focal length value to obtain a clear tracking picture.

4. The method for tracking and monitoring the multiple targets in the complex environment in real time according to claim 1, wherein the method for tracking and monitoring the targets by the video processing module comprises the following steps:

SS1, a first video acquisition module and a second video acquisition module are used as a group of video acquisition modules, a plurality of groups of video acquisition modules are arranged in a global monitoring area, and local monitoring areas of at least two adjacent groups of video acquisition modules are partially overlapped to ensure that the global monitoring area has no monitoring dead angle;

SS2, establishing a plane coordinate system by taking any point in the global monitoring area as an origin, and marking the overlapping area of at least two adjacent local monitoring areas;

SS3, a video acquisition module continuously tracks and shoots the tracked target in real time according to the path and the motion velocity vector of the tracked target and an image recognition algorithm, and the video acquisition module calculates the real-time coordinate information of the tracked target in the monitoring area of the video acquisition module;

SS4, when the tracked target enters the overlap area, the video processing module receives the video monitoring information of the local monitoring area related to the overlap area, predicts the advancing direction of the tracked target in the overlap area according to the velocity vector of the tracked target, the video acquisition module in the corresponding direction prepares to perform relay tracking, and when the video acquisition module in the predicted direction does not acquire the corresponding target outside the overlap area, the original video acquisition module performing target tracking continuously tracks.