Preprocessing And Tracking Method Of IR Small Target Under Complex Backgroud

Because the infrared target tracking technique has the advantages of high sensitiviy,high spatial resolution,antielectromagnetic interference ablity is strong,can identify the target in the complex background,tracking the target of many kinds,all-weather work time,can be achieved after launching,adapt to the characteristics of strong function,has become the focus of research around the world in military applications and key technology for development. However,due to the inherent characteristics of infrared imaging,there are many problems need to solve, infrared image preprocessing technology and the target tracking technology is studied.This paper first discusses the necessity of infrared image preprocessing technology,including blind pixels and non uniformity problems,these problems greatly affect the imaging quality of the system,reduces the detection distance of infrared system,directly restricts the final performance of the system.Non uniformity and blind pixels is the first problem to solve in the infrared focal plane array imaging system. This paper also discusses the reason and characteristic of non uniform and blind pixels,studied the application of several common processing method.Secondly,in the research of tracking technology, mainly studies the classical methods for IR small target tracking. They include the method based on Mean shift, the method based on particle filter that realize recursive Bayesian filter, the method based on max probability of observation, and the method based on probability data affiliation. The paper takes some characters of IR target based on target of IR image. The paper gives the probability of every character. They are applied above-mentioned methods. The new methods improve the robustness of tracking.Finally,the simulation experiment shows that the algorithm is better than the traditional algorithm,and made a summary of this paper and the prospect of future work.