新模型可实现慕士塔格全天相机云图自动分类

Core Viewpoint - The Xinjiang Astronomical Observatory has developed an automatic cloud classification model (ASCNet) to enhance the efficiency of sky condition assessment for astronomical observations, published in the journal "Research in Astronomy and Astrophysics" [1][2] Group 1: Technology Development - The ASCNet model utilizes artificial intelligence to enable machines to understand sky conditions, which is crucial for the operation of optical telescopes [1] - The model features a complementary dual-channel structure for feature extraction, capturing both global semantic information and local brightness texture characteristics for comprehensive cloud condition assessment [1] Group 2: Performance and Application - Testing results indicate that the model achieves a high consistency rate of approximately 92.7% with manual interpretation results, effectively identifying various typical cloud conditions [1] - The automatic classification of cloud images is significant for reducing manual workload and improving site monitoring efficiency, supporting the trend towards intelligent and refined astronomical observation [2]