Cloud Detection of Remote Sensing Image Based on Multi Feature Fusion

摘要

The existing cloud detection algorithms mainly include multi-channel threshold method and cloud detection method based on image features. The multi-channel threshold method is usually only for the reflectance of remote sensing image. The method based on image features is not enough for information mining of remote sensing image, and the feature expression ability is not strong enough, which results in the problems of low accuracy and poor robustness of the algorithm. In this paper, we propose to extract point features and block features from different forms of cloud to fuse multiple features, so as to achieve the features with good expression ability. Combined with active learning of support vector machine, more representative samples are fully selected for training, so as to achieve a higher classification accuracy rate in the case of fewer samples. The experimental results show that the cloud detection method proposed in this paper has high accuracy, good robustness and fast computing speed, so it can adapt to the application of satellite platform.