Addressing the issue that traditional change detection methods cannot or are difficult to identify types of changes, this book improves the change vector analysis method. According to the dominant features of each image object, different adaptive weights are assigned to each feature, constructing a multi-feature change detection theoretical framework based on adaptive weight change vector analysis. This framework mainly includes: ① calculation methods for change intensity and change direction; ② setting methods for adaptive weights; ③ expression of the algorithm in the polar coordinate system. Under this framework, a “two-step” detection pattern for multi-feature changes of high-resolution image objects has been developed. Firstly, the EM algorithm is used on the difference image to obtain the threshold of change intensity, achieving the extraction of changing objects; then, for these objects, the K-means method is used to distinguish different change directions, achieving the identification and differentiation of different types of changes.