从黄土高原植被指数数据中获取森林覆盖率信息

 由于存在既包含农作物又包含森林植被的混合像元,以年平均归一化植被指数(NDVI)作为反映森林覆盖率的指标有很大的局限性。在黄土高原,森林植被的季节变化与农作物的季节变化有很大差异,基于这种差异,以不同月份NDVI的不同组合和考虑某种权系数后计算出来的加权平均NDVI与森林覆盖率相联系,筛选对于森林覆盖率有较好反映的NDVI指标。结果表明:6、10月2月森林覆盖率与NDVI的决定系数最高,故以6、10月平均NDVI(INDVI,6+10)作为反映森林覆盖率的指标;全年森林覆盖率与INDVI,6+10之间的决定系数为0.5384,比全年森林覆盖率与全年平均NDVI之间的决定系数0.4432提高21.48%;以各月的决定系数作为权系数求得的NDVI加权平均值(INDVI,w)与全年森林覆盖率之间的决定系数为0.5144,比森林覆盖率与年平均NDVI(INDVI,m)之间的决定系数0.4432提高16.07%,因此,推荐全年各月加权平均NDVI(INDVI,w)或6、10月平均NDVI(INDVI,6+10)作为反映黄土高原森林覆盖率的遥感植被指标。

Extracting the information of forest coverage from Normal Difference Vegetation Index data of Loess Plateau

Due to the mixed pixels containing both crops and forest vegetation,there are limitations if the annual mean Normal Difference Vegetation Index(NDVI(INDVI,m)) is used as an index reflecting forest coverage.In the Loess plateau region,the seasonal variation of forest cover is quite different from that of crops.On this basis,the NDVI-based index of forest cover may be established by considering some combination of monthly NDVI or weighted average of monthly NDVI with some proper weight coefficients.It has bee...