期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2020
卷号：98
期号：14
页码：2755-2762
出版社：Journal of Theoretical and Applied
摘要：This work presents a system that combined well-known morphological filters in order to find true moving regions from a sequence of images. For localizing the changed region, a block-based change segmentation process is proposed. This change region is naturally a coarse region and also contains some holes. To compensate this, we used an edge-based dilation to get an anisotropic expansion of the coarse image. Then the segmentation is generated using watershed algorithm. To prevent over segmentation, we used a specially weighted gradient image to achieve segmentation. Also, we removed some local minima from that gradient image. Finally, a fusion is applied on morphological filters. Furthermore, we calculated the coverage ratio of edge pixels of each segmented region. Comparing with a converge threshold, we determined whether the segmented region is truly belongs to a moving region or not. In the end, the experimental result demonstrated the validity of our proposed method.
