森林扰动遥感影像检测方法研究进展

将碳达峰、碳中和纳入我国生态文明建设整体布局是中共中央作出的重大战略选择。森林是陆地生态系统最重要的贮碳库，对实现碳中和具有重要作用。森林的扰动与碳储量和碳汇关系密切，目前由于森林扰动资料的缺乏，导致区域森林碳汇与碳源的估算精度不准确，建立和完善适用于中国森林的扰动检测方法具有重要的理论意义和应用价值。文中在分析国内外森林扰动遥感检测文献的基础上对分类后比较法、直接分析法、时间序列分析法与深度学习法进行对比，分析每一类方法的技术特点和应用效果，指出森林扰动检测方法存在的问题，展望今后发展方向，以期能够为森林扰动检测相关研究提供参考。

Review on Forest Disturbance Detection Methods By Remote Sensing

Incorporating carbon peaking and carbon neutrality into the overall layout of building ecological civilization is a major strategic choice made by the Party Central Committee. Forests are the most important carbon reservoirs in terrestrial ecosystems and play an important role in achieving carbon neutrality. The disturbance of forests is closely related to carbon stock and carbon sequestration. At present, the shortage of forest disturbance information leads to inaccurate estimation accuracy of regional forest carbon sink and carbon source, and it is therefore of great theoretical significance and application value to establish and improve the disturbance detection method applicable to the forests in China. In this paper, we compare the forest disturbance detection methods based on literature review on forest disturbance detection by remote sensing. We also compare the disturbance detection methods, i.e., post-classification comparison method, time series analysis method, direct analysis method, and deep learning method, to analyze the technical characteristics and application effects of each method, and point out the problems found in forest disturbance detection methods. The future development of forest disturbance detection methods is discussed, to provide reference for the research in relation to forest disturbance detection.