基于NOAA/AVHRR数据的地表覆盖变化检测方法与监测

应用图像差分技术、NDVI斜率、生物量斜率计算了过去20年间(1981～2001)我国西部植被的变化状况，总结了计算结果的相关特征。统计表明三种方法所揭示的植被变化信息基本一致，而且前10年(1981～1991)与后10年(1991～2001)相比，植被生长状况呈负相关，并且前10年的植被生长状况要好于后10年。对比主成分分析结果。认为生物量斜率方法比图像差分方法、NDVI斜率方法较为准确，相对于主成分变换计算简明、生物物理意义明确，更适合于植被变化的检测和监测。在讨论卫星更替等外部因素对NDVI影响的基础上，对生物量斜率分割图像的统计揭示了我国西部在1981—2001的20年间的植被演化状况：林地、草地都发生了高比例、大面积的退化。只有很小比例的植被得以改善。表现出局部改善、总体恶化的局面。

Detection Methods And Land Cover Change Analysis Using NOAA /AVHRR NDVI Datasets

This paper analyzes the NDVI datasets of long time series and separately yields statistic results, using simple differencing, slope map of NDVI, slope map of biomass, and principle component analysis techniques. The correlation characteristics for the images of change acquired according to the algorithms mentioned above, implies that the majority pixels of the different land covers have experienced nearly the same changes in change direction and magnitude during the same period, the land covers of the former 10 years (1981-1991) are in a better growth than the latter 10 years (1991-2001). Comparing with the other three algorithms (simple differencing, slope map of NDVI and principle component analysis), the slope map of biomass method indicates more precise than the simple differencing and slope map of NDVI, and is more explicit in change analysis and more constructive in biomass physics. Based on discussion of the external factors affected the NDVI, like satellite shift or sensor degradation, the statistics on the slope images of biomass indicate that land covers deteriorated extensively in the past few years which confirmed such a fact that the status of the vegetation of the west part of China in the past 20 years (1981-2001) is suffering an extensively deterioration and only an improvement in some parts of the region.