及时准确的森林干扰信息是掌握森林时空演变规律和碳汇状态变化的重要基础知识.基于1999—2010年间的Landsat TM/ETM+遥感影像,本研究采用单类支持向量机(one class support vector machine,OCSVM)监测了近12 a以来西双版纳傣族自治州的热带森林干扰的时空变化.监测结果表明:1999—2010年间,西双版纳热带森林面积在各种因素的干扰下呈逐年减少趋势,热带森林保护形势依然严峻.OCSVM是一种快速提取热带森林面积的简单方法,具有较高的提取精度(90.19%).将长时间序列MODIS数据与Google Earth高分辨率影像结合是快速获取丰富的高质量采样数据的有效途径,能够为遥感分类算法训练和结果验证提供可靠数据源.%Up-to-date and accurate forest disturbance information is very important to understand temporal-spatial changes and carbon sinks in tropical regions. One class support vector machine ( OCSVM ) was used for monitoring forest disturbance of Xishuangbanna Dai Nationality Autonomous Prefecture based on Landsat TM/ETM+ images from 1999 to 2010 at 30 m spatial resolution. The result showed that: tropical forest area of Xishuangbanna appeared a yearly decrease trend disturbed by various factors during between 1999 and 2010, and the situation of tropical forest protection was still serious in the future. OCSVM performed a high accuracy (90.19%) in extracting tropical forests and was very simple and efficient. The combination of multiple temporal MODIS data and high resolution images of Google Earth is an efficient way to acquire abundant sampling datum with high quality which is able to provide reliable samples for algorithm training and result validation of remote sensing classification.
展开▼
