Abstract: In order to suppress complex background, an improved detection algorithm based on complex filter bank is presented for dim-small target under infrared background clutter. First, the spectrum characteristics of both the complex background image and the target image are analyzed, and the idea of sub-band processing is presented. Then after the comparison filters, the complex filter bank is adopted to decompose the infrared image with dim-small target into sections of different frequency bands. Next, the Robinson filter algorithm is used to process the decomposed sub-band images for extracting target. And finally these images are integrated with the method of weighting. The targets in the final processed image are enhanced notably and the backgrounds are suppressed effectively. The experiment