Clutter Infrared Background Suppression Technology Based On Adaptive Estimation

The detection problem of point targets in IR images is a hot spot nowadays. Many scholars have dipped into this field and put forward valuable methods, such as image sequence detection and motion tracking. However, the computation load is too heavy to meet the practical engineering. According to this problem, this paper addresses the issue of small target detection with background estimation technology. As well known, the SNR of the infrared image is very low for the clutter background takes over most of the image. It's very difficult to detect the dim target in such condition. Thus, in order to detect the weak targets from clutter background, the background suppression of infrared image is necessary. Firstly, the thermal imaging theory and the characteristics of infrared image are analyzed. This paper studies the background suppression based on the adaptive filtering technology. Particularly saying, we analyze three kind of filter: Linear adaptive filter, nonlinear mean filter and neural network predictor. In addition, we present improved algorithms. Experiments are done to analyze the performance. The results show that the adaptive background technology performs well in background estimation. Background is suppressed after filtering. Moreover, it is effective and time-saving with practical value in engineering.