Abstract: Due to the disadvantage of timcconsuming and large volumes data of video processing, based on the hypothesis that the background pixel intensity always appears in image sequence with maximum probability, a novel method of moving objects segmentation was presented. This approach extracts the approximate components of video images sequcnce on hand firstly using the DWT (discrete wavclet transform) , and then selects the higher frequency pixel values of an interval to reconstruct the background according to the pixels clustering combing dual-threshold with similar classes merged method, at last verifies background reconstruction accuracy by quoting the evaluation criteria of image matches. Experimental results demonstrate that the proposed method can extract better background to achieve efficient and integrated moving object segmentation.

Key words: DWT, Background subtraction, Pixels clustering, Background reconstruction, Moving segmentation