Cloud and Snow Segmentation in Satellite Images Using an Encoder–Decoder Deep Convolutional Neural Networks

The segmentation of cloud and snow in satellite images is a key step for subsequent image analysis, interpretation, other applications. In this paper, method based on deep convolutional neural network (DCNN) with enhanced encoder–decoder architecture—ED-CNN—is proposed. method, the atrous spatial pyramid pooling (ASPP) module used to enhance encoder, while decoder fusion features from different stages which improves accuracy. Comparative experiments show that proposed superior DeepLabV3+ Xception ResNet50. Additionally, rough-labeled dataset containing 23,520 fine-labeled data consisting 310 TH-1 are created, where we studied relationship between quality quantity labels performance segmentation. Through same datasets, found related more closely rather than their quality. Namely, under labeling consumption, using only performs better plus 10% images.