The demand of target information acquisition and application is becoming more and more urgent all over the world, and all countries are focusing on the research of new target information acquisition technology, especially for all kinds of targets, especially for air targets. In today's information society, the new information acquisition technology is of great significance, which can promote social development and improve people's living standards, and also play a significant role in improving the national defense system and ensuring national security. Therefore, it is necessary to study the new target information acquisition technology. Target detection technology based on atmospheric disturbance is a new information acquisition technology system. The use of atmospheric disturbance in target detection is not affected by the performance of the target itself, and has great application potential.In this paper, four main image processing algorithms of target atmospheric disturbance detection are introduced, which are cross-correlation method, optical flow method, frame difference method and background detection method. The technical principle is described and the technical advantages and disadvantages are analyzed (Tab.4). The cross-correlation algorithm has good real-time performance, but will reduce the resolution; The optical flow method has high precision but poor real-time performance. Inter-frame difference method has good real-time performance, but poor accuracy and applicability. Background detection method has good accuracy and poor applicability. According to the development status of the four algorithms at home and abroad, through comprehensive research, the optimization methods of various algorithms are analyzed and summarized, which can be divided into three categories (Tab.2) of optimization algorithm itself, integration with other image processing algorithms, and neural network based. Through the analysis of relevant literature, the advantages and disadvantages of the three algorithm optimization methods are revealed (Tab.3). The optimization algorithm itself has low complexity and can achieve high real-time performance, but limited by the basic principles of the algorithm, the optimization effect is not obvious; The method of fusion with other image processing algorithms can make up for the technical limitations of the algorithm, achieve high performance and high robustness, but the complexity of the algorithm increases, and the real-time performance is affected. The optimization method based on neural network can greatly improve the algorithm performance and achieve high adaptability, but it requires a lot of prior information and has poor real-time performance. Based on this, the four methods and the future development direction of target atmospheric disturbance image processing are prospected and summarized.Based on the analysis and summary of the research progress of target atmospheric disturbance image processing methods at home and abroad, the optimization method of image processing algorithm in target atmospheric disturbance detection is given as follows. At present, it is necessary to develop the technical direction of optimizing the image processing methods of atmospheric disturbance target detection by using other algorithms, such as inter-frame difference method combined with optical flow method, and overall optimization of multiple target atmospheric disturbance algorithms by using machine learning technology. Facing the future, the image processing method of atmospheric disturbance target detection based on small sample unsupervised learning has great application prospect.