Automatic video segmentation, which is based on SBD (shot boundary detection), is one necessary step in establishing video databases. Existing SBD algorithms can detect abrupt scene changes quite accurately but may fail for gradual scene changes. This is due to the fact that there are no apparent peak values of frame difference during gradual scene change periods and hence it is more difficult to detect them. Wipes are frequently used to edit video in spatial domain to generate various kinds of scene changes. Through analysis of various wipes, a new algorithm is proposed in this paper based on a spatial editing model for wipe scene change detection. This method is better than the statistic-based method proposed by Alattar. Test results for video sequences generated by Adobe Premiere 5.1 containing wipe scene changes show that the algorithm is well suited for various types and durations of wipes.