Supervised Cloud Presence Detection in Remote Sensing Image

We propose a supervised method of judging the presence or absence of clouds in remote sensing images. Cloud detection is the basis for further removal of cloud occlusion and landmark detection. We establishes a data set related to clouds using the remote sensing image data from the landsat8 satellite. Next, we build a convolutional neural network for the task. Deep learning method for cloud detection is suitable for different forms of clouds. The detection method reduces the features extracted by the deep network to a one-dimensional space and compare the features with the label information to train the network. Finally, we evaluated the effectiveness of this method in cloud detection.