Detection And Recognition And Thread Assessment For Small Target At Low Altitude And Slow Speed

Detection and recognition of small targets at low altitude has been a difficult problem for low-altitude detection systems.With the opening of low-altitude airspace in recent years,it is urgent to solve the problem of detecting and identifying low-speed and small targets.In this paper,based on the analysis of imaging characteristics,motion characteristics and self-characteristics of low-speed slow-moving small targets,a low-slow detection and recognition and threat assessment method is proposed.The aim of this paper is to study the low and slow targets of long-range high-field photoelectric detection equipment,mainly to solve the detection and recognition of civil low-altitude aircraft.In this paper,three aspects are studied: detection,recognition and threat degree evaluation.The main contents are as follows:First,target detection is a prerequisite for target recognition and threat assessment.Aiming at the complicated imaging background and large range of the field of view,a method of threshold segmentation,morphological processing and edge detection is used to preprocess,and then the hough transform is used to detect the image.It is divided into two parts: the sky and the urban background.Remove the complex low-level background of the city and then extract the moving target only in the sky area with the method of block frame difference and the morphological filtering.Secondly,on the basis of accurately extracting the moving target,a target recognition method combining the global characteristic and local characteristic of target is proposed according to the different characteristics of different targets.This method first extract the four target features of Hu moment,affine invariant moment,rectangle detection and motion trajectory and then normalize them use the maximum and minimum method.Finally the SVM support vector machine classification method is used to classify low-altitude targets.Then,according to the result of target recognition,the six factors,such as target category,target speed,distance from important places,target armed performance,target queryability and target to be protected,are weighted as threat assessment factors.The threat degree of sorting is evaluated by the combination of fuzzy sets and information entropy and TOPSIS.Finally,based on the experimental platform of VC ++ and opencv computer vision library,the experimental algorithm is experimentally verified,and the experimental results are summarized and analyzed.