Multi-camera moving target continuous tracking method for mining

Abstract

The invention provides a multi-camera moving target continuous tracking method for mining. The method mainly includes three major systems: the establishment of a dictionary for target templates, target detection, and sparse representation tracking. During tracking, all cameras are mounted in fixed positions. The core is that a target detecting and tracking algorithm is adopted, moving targets of interest are searched from a large amount of video data, a computer is used to replace manual monitoring, the time for manual search is reduced, and the efficiency of search is improved. The route of underground workers is provided for ground management personnel, so that the management personnel can inform the underground workers of the change of the surrounding environment in time and warn the underground workers about the danger in the surrounding environment in time. A detecting and tracking algorithm between different cameras is put forward, and the approximate position of the staff can be known. When a mine accident occurs, a rescue target can be found in time, so that time can be bought for the rescue team and rescue can be implemented in time.

Description

technical field [0001] The present invention is in the field of moving target tracking, in particular to a method for continuous tracking of moving targets with multiple cameras in mines. Background technique [0002] At this stage, the monitoring system of coal mines is still mainly based on manual monitoring, but scientific research shows that if the human eye continuously monitors the screen for more than 20 minutes, visual fatigue will occur, resulting in cerebral palsy, and various emergencies cannot be dealt with in real time. Therefore, it is necessary to introduce intelligent vision technology into the coal mine video monitoring system, and use machine processing to replace human eye observation, thereby improving the efficiency of monitoring. At present, general coal mines are equipped with monitoring systems, and these measures can avoid some accidents to a certain extent or provide analysis data required for rescue after an accident occurs. However, once the coal...

AI Technical Summary

Problems solved by technology

[0004] (1) The environment under the mine is relatively dark, and artificial lighting is used all the year round. Although there are lighting devices, this lighting method is different from natural light, resulting in uneven lighting intensity in the coal mine, even in the same monitoring environment. Under different conditions, the distance from the light source will also lead to different lighting

[0005] (2) In the mine environment, except for some special signs or equipment that are eye-catching and easy to identify and process, the image information in other areas is almost mainly in black, white, and gray. Not very rich color information can be exploited

[0006] (3) Under normal circumstances, coal mines will equip each worker with a miner's lamp. When a worker enters the detection area with a miner's lamp, it will have a great impact on the brightness of the area

[0007] In order to solve the above problems, the prior art discloses some tracking methods of moving objects, such as CN104091350 discloses an object tracking method using motion blur information, which uses sparse representation to solve the problem of image degradation caused by motion blur, and uses the extracted Motion information, combined with the particle filter algorithm, makes the tracking algorithm more effective; but this method cannot solve the problem of continuous tracking of targets between different fixed cameras in the mine environment, and the video environment of the coal mine has uneven illumination and lack of color information. The problem that the target is close to the background environment leads to poor application of traditional target detection and tracking algorithms in mines, so there are many cases where target tracking is lost

Images