A detection method for multi-type earth's surface anomalies based on multi-dimensional feature space

On-orbit processing is an important research direction in the real-time remote sensing detection of earth's surface anomalies (ESSA). However, existing methods cannot comprehensively utilize multi-dimensional remote sensing characteristics to detect multi-type ESSA simultaneously. Meanwhile, it is difficult to realize the comprehensive utilization of multi-dimensional remote sensing characteristics for the limited storage and computing resources on satellites. Therefore, this study proposed a detection method for multi-type ESSA based on multi-dimensional feature space. The proposed method first selected the remote sensing characteristics reflecting the basic earth's surface elements to construct a multi-dimensional feature space and proposed two comprehensive remote sensing characteristics. Then, these characteristics were used to build a prior knowledge base reflecting the normal earth's surface conditions. Finally, by comparing the real-time acquired data and prior knowledge base, this study completed ESSA detection. The validation results indicated that the proposed method can effectively detect multi-type ESSA with an accuracy of over 85%. Moreover, the proposed method simplifies the large and complex ESSA remote sensing characteristic system, which greatly reduces the complexity of ESSA detection methods and increases the possibility of on-orbit processing.