基于NDVI象元二分法的植被覆盖变化监测

Monitoring vegetation-cover changes based on NDVI dimidiate pixel model

研究了NDVI象元二分法计算植被覆盖度的精度,并对其适用范围进行了分析。利用2002年、2007年合肥市的TM数据,针对于城市地区的特点,选择合适的NDVIsoil和NDVIveg,计算各自时相的植被覆盖度,然后采用差值运算方法得到该时段合肥市区植被覆盖度变化数据。对于合肥市来说,在研究时段,植被覆盖度总体呈下降的趋势,下降最为明显的是滨湖新区和包河工业区,下降值分别为16.7%和21.4%,植被覆盖的锐减是由于城市扩张造成的,是一个值得注意的生态问题。与航片调绘结果进行比较,NDVI二分法得到的植被覆盖度对于低植被覆盖区偏高,对于高植被覆盖的林区偏低,两个时相植被覆盖度差值可以消除部分系统误差的影响,具有更高的精度,可以作为植被覆盖监测的重要手段。

The accuracy and scope of vegetation fraction by NDVI dimidiate pixel model was studied in this paper. Selecting the appropriate NDVIsoil and NDVIveg for urban areas, vegetation fraction values were calculated respectively based on TM of Hefei City, Anhui Province in 2002 and 2007 by NDVI dimidiate pixel model, changes in vegetation fraction were obtained by image – minus method. For Hefei City, during the study period, vegetation coverage in general decreased. The most obvious was the decline in Binhu New Area and Baohe Industrial Park, decreased values were 16.7% and 21.4%, respectively. The sharp reduction of vegetation cover caused by urban expansion is a remarkable ecological problem. Compared with the results of aerophotograph surveying, the value of vegetation fraction obtained by this method is higher in low vegetation–covered area, and which is lower in high vegetation–covered forest area, but the vegetation fraction change calculated by image –minus method can eliminate some system error and achieve higher precision in all areas, so image –minus method based on NDVI dimidiate pixel model can serve as an important means of monitoring vegetation cover.