Improved Background Prediction Algorithm for IR Small Targets Detection

For detection problem of infrared dim small targets in complex background, an adaptive filtering algorithm based on the improved M-estimation is put forward to suppress background clutter. In the proposed algorithm, target pixels and observed noise form the mixed interference of background estimation. In order to better estimate the background, the proposed algorithm introduces the correction factor to reduce the influence of target pixels and introduces the forget factor to entreat the non-homogeneous backgrounds. The experimental results to real images show that the proposed algorithm can better suppress background and preserve target information than the commonly-used median filtering algorithm and the LMS filtering algorithm. As a result, the proposed algorithm can detect effectively IR small targets in complex background.