关键词: 高寒湿地, 高分卫星数据, 分类, 麦地卡, 监测模型

Abstract: The unique alpine wetlands of the northern Tibetan Plateau provide advantageous resources and environment for Tibet and the whole country. Wetland degradation is becoming more and more serious due to human activities and natural causes. So far, this paper uses GF-1 remote sensing data to conduct multiple experiments to complete the establishment of a hierarchical classification decision tree, comprehensively considers the spectral characteristics and texture characteristics of wetlands, and combines appropriate scale segmentation images and various major wetland information identification methods. Finally, an alpine wetland monitoring model based on high-resolution satellite data is established for the northern Tibetan Plateau. This model can realize automatic extraction and classification of wetland information. The results lay a foundation for the research on wetland degradation and its ecological restoration. Using the model to monitor the wetlands of the Mcdika Nature Reserve in 2021, the total area of wetlands is 319.02 km², and the wetland area accounts for 36.26% of the total area of nature reserves. The areas of various types of wetlands from large to small are herbaceous swamps, lake wetlands, river wetlands, glacial snow, peat swamps, and floodplain wetlands. By randomly selecting test points and adopting the method of confusion matrix, the monitoring accuracy is evaluated. It is found that the total classification accuracy is 86.83%, and the classification accuracy Kappa coefficient is 0.827 5, so the model has achieved a good use effect.

Key words: alpine wetlands, high-resolution satellite data, classification, Mcdika, monitoring model