基于Landsat8植被覆盖度反演及应用——以宜黄县为例

Inversion and Application of Vegetation Coverage Based on Landsat8——Set Yihuang County as an Example

植被覆盖度是反映地表植被状况和衡量生态坏境的重要标志之一。利用2016年宜黄县域范围的Landsat8 OLI影像,在对研究区域进行土地利用分类时,采取最大似然法和基于CART算法进行监督分类和决策树分类。通过比较两者的统计数据,得出基于CART算法的决策树分类精度要更高。然后利用决策树分类所得的土地利用图采用归一化植被指数(NDVI)和改进的像元二分模型反演出宜黄县植被覆盖度,最后通过宜黄县实际植被覆盖率与反演结果进行对比,并对其定性分析。

Vegetation coverage is one of the important symbols which reflect the aboveground vegeta- tion status and measure the ecological environment. Landsat80LI image in range of Yihuang County in 2016 was used in this paper. When the land use classification was carried out against the study ar- ea, the maximum likelihood method and CART-based algorithm were used in this paper to carry out the supervised classification and decision tree classification. By comparing the statistical data of both methods, it was concluded that the accuracy of decision tree classification based on CART algorithm was much higher. Afterwards, the land use map obtained by the decision tree classification was used, and the normalized difference vegetation index （NDVI） and the improved pixel binary model were applied to invert the vegetation coverage of Yihuang County. Finally, the actual vegetation coverage and inversion result of Yihuang County were compared, and the qualitative analysis was carried out.