 The cloud detection model, CDM, proposed in this paper uses ResNet 18 as its backbone to extract features from images at various levels. It then uses the multi-scale global attention module to enhance the channel and spatial information, which improves the accuracy of detection. Additionally, the strip pyramid channel attention module is used to learn spatial information at multiple scales, allowing for the detection of smaller clouds more accurately. Finally, the CDM combines the high-dimensional feature and low-dimensional feature using the higher-archical feature aggregation module, followed by upsampling layer by layer. This produces a more accurate segmentation result than existing methods. This article was authored by Chao Zhang, Li Guo Wang, Li Ding, and others.