Abstract: A new detection algorithm for a moving dim target in the dual-band infrared image based on fuzzy fusion is proposed. The algorithm, according to the noise characteristics of the differential images obtained from dual-band IR images, introduces the membership function that evaluates the intensity of gray-grade variation on each pixel point of the differential image, and then several differential images are fused with the membership function. The dim target’ possible trajectory in single-band IR images is found after the result image is dealt with by the operation of fuzzy reasoning according to the rule of ‘IF…THEN…’, and the final result should be obtained from the trajectory fusion of dual-band IR images. Experimental results show that our algorithm avoids the shortcoming of low detection probability caused by difficult threshold determination and “hard” decision in terms of fuzzy reasoning, improves the detection performances by fusion of dual-band IR images, and effectively detects the dim moving target.

Key words: dual-band IR, dim target detection, fuzzy membership, fuzzy fusion, IR image sequence