Abstract: A change detection method based on sequential decision fusion is proposed by improving the decision tree with sequential analysis. Providing that decision tree algorithm can be used to process both continuous data and discrete data, an iterative process is designed to generate new input variables, which are composed of the previous input variables and the classification results. The decision tree is trained by the iterative process until the output variables are stable. The proposed method can reduce the uncertainty in the decision-tree-based change detection method. Experiments prove the feasibility and effectiveness of this method, and the results show that it provides a new approach to improve the accuracy of change detection for remote sensing images.

Key words: decision tree, sequential analysis, decision fusion, change detection