基于多特征融合的面向对象冰川边界提取

Pixel-based classification struggles with the accurate identification of glacier changes in areas with similar spectral characteristics, particularly in debris-covered areas where spectral features closely resemble the surrounding mountains and rocks, thereby resulting in low extraction accuracy. This study investigates the Yinsugaiti and Yalong Glaciers using Google Earth Engine to integrate spectral indices, microwave texture features, and topographic data. An object-based(OB) machine learning algorithm is applied for automated glacier extraction and compared to pixel-based(PB) classification methods. The results show the following. (1) The OB classification approach, integrating multi-feature fusion, significantly improved the glacier extraction accuracy. The OB_RF classifier achieved an overall accuracy of 98.1%, a Kappa coefficient of 0.97, and an F1-score of 98.67%, outperforming the OB_CART and OB_GTB classifiers. When compared to PB_RF, the overall accuracy, Kappa coefficient, and F1-score increased by 1.7%, 0.024, and 5.57%, respectively. (2) Between 20012022, the Yinsugaiti and Yalong Glaciers retreated at average annual rates of 0.08% and 0.13%, respectively. (3) Supraglacial debris was primarily distributed below 5,000 and 4,800 m on the Yinsugaiti and Yalong Glacier, respectively. Over the same period, debris-covered areas on both glaciers expanded upward.