Design And Implementation Of Optoelectronic Tracking System Based On Video Detection

In recent years,with the gradual liberalization of low-altitude airspace control and the rapid development of "low-slow-small" aircraft,the use of "low-slow-small" aircraft for illegal flight and terrorist attacks has been on the rise.Such events have disrupted normal airline order and caused big economic losses and negative effects.Therefore,reasonable and effective detection and tracking of "low-slow-small" aircraft has become a research hotspot in the field of moving target detection and tracking.This paper designs and implements an electro-optical tracking system based on video detection.The system can accurately detect “low,slow,and small” drones with a detection range of up to 800 meters.Firstly,combing and summarizing the classic target detection and tracking algorithm.The detection algorithms include background difference method and inter-frame difference method.The tracking algorithms include Mean-shift algorithm and particle filter algorithm.The algorithm is applied to detection,tracking and tracking of low-slow and small-drone UAVs.The results show that the background-difference method has high dependence on the background,which will reduce the detection accuracy of moving targets and cause judgment errors.The inter-frame difference method can detect the basic overview of the moving target,but it is difficult to choose the binarization threshold of the algorithm.In addition,when the target speed changes,it will cause the phenomenon of “ghosting” and “hole” in the detection,resulting in false detection of the target.Mean-shift tracking algorithm is simple in form and high in realizability.It can be tracked in real time when the target tracking area is known.However,this algorithm has low adaptability to environmental changes,and the particle filter tracking algorithm has strict restrictions on the use of conditions and loses the target.The probability is high and lacks wide applicability.Then,a target tracking detection algorithm based on 9₇ lifting wavelet and region growing is proposed to analyze the image information from the spatial domain to the frequency domain.The algorithm first carries out the 9₇ lifting wavelet transform on the image,and enhances the high-frequency and low-frequency contrasts of the target and the image background.Since the shadow of the moving target may cause the background local screen brightness to change,threshold filtering is used to remove the ambiguity target caused by this change.The region growing algorithm divides the threshold image into different connected regions and marks the target region.The target is judged according to the number of pixels in each region and the geometric characteristics of the “low-slow-small” drone to be detected in this paper.Then,remove the misjudgment caused by the “spot” and “peak”,determine the target position in the single frame image,associated with a few frames of a single frame image to remove the misdetection of multiple targets such as "bird",and finally find the target area and output position.Finally,the system was integrated and tested on the optoelectronic tracking system.The system was divided into four modules according to the function.1)The video image acquisition module,completes the image acquisition of the target and uses the network to transmit the detection and tracking module;2)The detection and tracking algorithm processing module,uses the 9₇ lifting wavelet and the regional growth algorithm to detect tracking drone;3)The turntable control module,controls the real-time tracking of the target according to the target position detected by the algorithm and controls the rotation of the two-dimensional turntable;4)Monitors the display module,In the video window of the software,the target motion trajectory in the scene is seen,and the user’s monitoring of the specific scene is realized.A large number of field tests were performed on the system and optimized.The system meets the research objectives of the project.