It is very important to use infrared imaging for target automatic detection in the modern defense technology. In this paper, all kinds of common segmentation algorithms of moving targets which can be used in the shadow compensation phase of target detection of infrared warning systems are further studied. The effectiveness of each algorithm in processing images is simulated by using a software. The advantages and disadvantages of different algarithms are summarized. A 2D maximum entropy approach based on the subtraction between two consecutive frames is proposed. The experimental results have shown that the proposed approach is effective for the detection of regional moving targets in infrared image sequences.