A relaxation segmentation algorithm of forward-looking infrared (FLIR) images with edge constraint is presented.An evaluation criterion based on distinct edge pixels is used to examine the segmentation results by the relaxation segmentation algorithm under different initial assignment of probabilities.Thus the good segmentation result can be achieved by automatically adapting initial assignment of probabilities to reach the optimal or suboptimal solution of the evaluation criterion.Experimental results with real FLIR images are given.