Abstract: The distance image means the one formed by distance measuring based on pulse rejection combined with beam scanning, whose statistical property is usually different from that of the conventional optical image. This paper proposes a hierarchical filtering algorithm for processing distance images, which can, compared with the conventional methods, remove noise more effectively and obtain images with clearer objects and background. It segments an image by taking advantage of the static characteristic of a distance image and filtering respectively, and then synthesizes to reconstitute a new image. Experimental results suggest that the algorithm behaves well and is suitable for DSP implementation in real time.

Key words: distance image, hierarchical filtering, image processing