基于NDVI像元二分模型植被覆盖度反演喀斯特石漠化研究——以贵州毕节鸭池示范区为例

Study on Karst Rock-desertification of Extracting Vegetation Coverage Inversion Based on NDVI Serial Images and Dimidiate Pixel Model

该研究在"3S"技术发展的背景下,以贵州省毕节鸭池示范区为研究区,选用2005年SPOT遥感影像,对遥感影像进行几何校正、图像配准等数据预处理,引入NDVI植被指数和像元二分模型,根据像元二分模型的理论,确定出NDVIsoil和NDVIveg,建立了反演植被盖度的像元二分模型。并根据大量的外业调查数据对所建立的模型进行了精度验证,定量估算了毕节鸭池示范区的植被覆盖度,取得NDVI植被指数与喀斯特石漠化之间的对应关系,得到研究区NDVI图像,最后由各覆盖度得出NDVI结果反演毕节鸭池示范区石漠化特征,得到研究区的各等级喀斯特石漠化的面积以及空间分布情况。利用该方法进行喀斯特石漠化研究,明显改善了分类的精度,为喀斯特石漠化生态修复及其他综合防治提供科学依据。

Under the development background of ＂3S＂ technology,taking Yachi demonstration area of Guizhou as the study region,SPOT images in 2005 were selected to pretreat the remote-sensing images including geometric correction and image registration.NDVI vegetation index and dimidiate pixel model were introduced to determine NDVIsoil and NDVveg according to the theory of the dimidiate pixel model.The dimidiate pixel model for the vegetation coverage inversion was established.Based on a large amount of filed survey data,the precision of the established model was validated.The vegetation coverage of Yachi demonstration area was quantitively estimated to analyze the relationship between NDVI vegetation index and karst rock-desertification and obtain NDVI images in the study area.Finally,the karst rocky desertification characteristics of Yachi demonstration area from NDVI results inversion were obtained according to each coverage.The area and spatial distribution situations of karst rock-desertification in the study area at different levels were obtained.The rock-desertification was studied by using this method,which improved the precision of the classification and provide the scientific basis for the ecological restoration and other comprehensive control of karst rock-desertification.