An Improved Fuzzy Clustering Method to Detect Moving Objects

The classical fuzzy clustering method needs to determine the number of group for classification before all samples are processed and the number of group is fixed during iteration, which dose not help to ensure the classification precision. Considering this, an improved fuzzy clustering method with elastic grouping logic is proposed. The elastic grouping logic, based on the samples´ ascriptions and their distances to the centers of each group, can dynamically adjust the number of group and achieve the accurate classification. Our improved clustering method is applied in the optical flow field. The experimental results show that our method has superiority over the classical clustering method in precision and can detect the moving object with precision.