Research On Airspace Target Detection Algorithm

In recent years,with the rapid development of science and technology and economy,the gradual rise of aviation trade and the widespread application of UAVs,the airspace environment has become increasingly complex.Aviation accidents have occurred frequently and ensuring airspace safety has become the top priority of current flight.The detection of flying targets in the air is the basic premise of maintaining airspace safety.It is of great significance to strengthen the management of airspace aircraft and maintain flight safety.This thesis conducts in-depth research on the detection algorithm of airspace moving target.Its main research contents are as follows:(1)For the disadvantage of The traditional Canny operator that the threshold needs to be manually set,an Otsu adaptive threshold method was proposed,which uses the Otsu method to select the best threshold and reduce the existence of false edges,so that the detected edges are more continuous and complete.(2)For the lack of edge contours of moving objects in the traditional frame difference method,a frame difference method based on edge features was proposed.The method makes the detection result of the target more complete by fusing the edge information of the target,and at the same time,it has a better suppression effect on the illumination change.(3)For the shortcomings of traditional Vi Be algorithm such as ghost elimination and loss of intermittent targets,a target detection algorithm that combines edge features,frame difference method and Vi Be was proposed.The algorithm expands the 8-neighborhood background sampling range to 24-neighborhood to improve the authenticity of the constructed background;uses the improved Otsu method to adaptively adjust the pixel classification threshold to improve the classification accuracy;combined with the Vi Be algorithm of frame difference,It could eliminated ghost quickly;the selective update of the background area reduces the complexity of the algorithm and ensures the continuous detection of intermittent moving targets;the use of contrast to screen suspected targets again eliminates cloud interference and improves target detection.Compared with the Vi Be algorithm,the detection effect of the algorithm in this paper has been greatly improved,the comprehensive index F has increased by 6.75%,and the overall performance index PCC has increased by 8.34%.(4)Aiming at the above improved algorithm,the high-level synthesis tool is used to transplant the algorithm to hardware,and a relatively complete target detection system is built,and the validity and reliability of the target detection system are verified by experimental tests.