Research On Infrared Small Moving Target Detection And Tracking Under Complicated Background

Infrared imaging guidance weapon plays an important role in the field of precision guided weapons for its advantages,such as high precision guidance,strong autonomy,strong anti-interference ability and all-weather operations.As the key technology of infrared imaging guidance weapon,the technology of infrared moving small target detection and tracking under complex background is a challenging research subject,and it has important theoretical significance and application value for its further research.Based on the development of strapdown infrared imaging seeker,in order to hit the moving ground target,the thesis studies the technology of small target detection and tracking under complex background.The main work and innovation are as follows:(1).A novel infrared small target detection algorithm based on frequency domain saliency is proposed.Since it is difficult for the conventional single-frame infrared small target detection algorithm to suppress the global repetitive clutter,from the the perspective of frequency domain saliency,the proposed algorithm can suppress global duplicate clutter and highlights small targets by suppressing the spike of the amplitude spectrum of the infrared image.Experimental results on different infrared images show that the proposed algorithm can improve the detection accuracy and anti-noise ability,also can effectively suppress the global repetitive clutter.(2).An infrared small target tracking algorithm based on Kernelized Correlation Filters(KCF)and distribution field is proposed.In order to meet the real-time and robust requirements of infrared small target tracking algorithm under the condition of missile loading,this paper introduces KCF tracking algorithm into infrared small target tracking field for the first time.In order to overcome the shortcomings of KCF tracking algorithm in tracking infrared small targets,the following three aspects are improved: Firstly,to enhance the robustness and uniqueness of the target expression,the small infrared target model is constructed by the distribution field.As a blurred image block which combines geometric information and gray information,distributed field is suitable for describing infrared small targets with only gray information and obvious target deformation.Then,Kalman filter can reduce the target search range and improve the speed of the tracking algorithm.Lastly,since the small target's shape changes greatly in sequence infrared images,a new model updating strategy is proposed to improve the stability and adaptability of the tracking algorithm.Experimental results show that the proposed tracking method based on KCF and distribution field is superior to traditional tracking algorithm in tracking precision and speed.(3).An infrared moving small target detection algorithm based on line matching is proposed.Since traditional detection algorithm has little effect on small infrared target under complex dynamic background,this paper firstly introduces the idea of eliminating background motion by line matching to the moving target detection field.Firstly,we extract the candidate target in the first frame by using the single frame detection algorithm in Chapter 2.Secondly,the corresponding position of the candidate object in the final frame image is obtained by the tracking algorithm in Chapter 3.Then,the purpose of the line matching between the first frame and the last frame in the detection sequence is to obtain the intersection points of the matching line.The RANSAC algorithm is used to obtain the transformation matrix to eliminate the global motion of the camera.Finally,Otsu algorithm is used to distinguish the moving target from the background interference based on the position offset error.The experimental results show that the detection algorithm proposed in this paper can adapt to the dynamic background with rotation,translation and scale change,and the detection effect is obviously superior to the traditional algorithm.