Research On Image Processing Algorithm Of TV Guided Ammunition

In modern warfare,the use of precision guided missiles has a very important influence on the direction of war.Therefore,it is of great significance to study precision guidance technology.Compared with other guidance methods,TV guidance not only has high concealment,but also has strong anti-interference ability and low cost.Image processing technology is the core of TV guidance technology.This article makes in-depth research on the key parts of TV guidance image processing,including image preprocessing,image matching,moving target tracking.First of all,the basic theories required for the research of image processing algorithms for TV guidance are introduced.The main theories are color space model,image grayscale transformation and image morphology operations,which will establish the foundation for subsequent research.Secondly,due to various interferences,the captured image of the bomb-mounted TV seeker will generate a lot of noise and degradation.In response to this,common methods of image pre-processing such as image denoising and image sharpening are studied.Focused on the research of image enhancement of TV guidance under severe weather,and the two common image enhancement algorithms of homomorphic filtering and Retinex theory are analyzed.In order to solve the lack of details of the former algorithm,this paper integrates the wavelet transform into the algorithm and adds CLAHE processing.The experimental results show that the improved algorithm can effectively enhance the quality of the image collected in bad weather.Furthermore,the performance of the image matching algorithm determines the overall performance of the TV guidance and is also the research basis of target tracking.This article describes the image matching theory based on gray information and corner features,and analyzes the principles,advantages and disadvantages of the SUSAN corners,Harris corners,and SIFT feature points.Image matching based on SIFT features is more suitable for TV guidance,but the SIFT algorithm is more complex and computationally intensive,which cannot achieve real-time matching.This paper proposes an improved SIFT algorithm based on edges.Only feature points are detected in the edge neighborhood.The Euclidean distance is replaced by the Mahalanobis distance as the criterion for feature point matching.Finally,the RANSAC algorithm is used to filter out the error correspondences of matching stage.The experiment shows that the time consumption of the improved algorithm is greatly reduced,and the accuracy is also improvedFinally,the tracking algorithm of TV guidance target is studied.The theory,advantages and disadvantages of CamShift algorithm and Kalman filter are analyzed.Aiming at the defect that the former algorithm will track failure when the target is blocked by other objects,this paper uses Kalman filtering to add the estimation of the target's future situation to make up for the above defects,and experimental verification.