有效识别采煤沉陷耕地损毁程度、面积及空间分布信息是矿区沉陷土地复垦、赔偿的客观需求。农作物受损后叶片叶绿素是很好的指示性指标，针对高潜水位矿区开采沉陷导致地面积水所引起的农作物渍害影响，基于无人机多光谱影像，在传统植被指数的基础上引入红边波段进行扩展，改进了现有的植被指数，结合田间同步实测叶绿素数据，采用经验模型法分别构建了单变量和多变量叶绿素反演模型，通过决定系数(R2)、均方根误差(RMSE)和估测精度(EA)3 个指标筛选出最佳模型。最终构建了多元线性回归MVI(red-edge),GNDVI,NDVI的叶片叶绿素定量反演模型，并采用野外对照区域样本对反演结果进行精度验证。结果显示:相对传统植被指数，引入红边后植被指数与叶片叶绿素的相关性得到显著增强，采用上述方法构建的多元线性回归模型的决定系数普遍提高0.10~0.20，达到了0.73，均方根误差降低了0.11~1.98，为0.938 SPAD，估测精度EA最终可达到83.4%，说明红边波段对采煤沉陷区作物的叶绿素响应敏感，引入红边波段构建叶绿素反演模型，可以应用于采煤沉陷影响下的玉米叶片叶绿素无人机遥感反演。结果显示:采煤沉陷盆地内玉米叶绿素含量主要集中于52~61 SPAD，面积占整体的75.1%，叶绿素含量在9~52 SPAD的作物面积占5.7%，玉米整体长势受采煤扰动影响较严重，叶绿素含量呈现从沉陷盆地边缘往中心逐渐降低的趋势。该研究为土地损毁监测与评价，土地复垦与生态修复等提供基础数据与理论支撑。

Identif the extent,area,and spatial distribution information of cultivated land subsidence damage is an ob- jective demand for land reclamation and compensation in mining areas. Leaf chlorophyll is a good indicator when crops are damaged This paper focused on the impact of crop waterlogging caused by mining subsidence on high groundwater mining Dong Tan coalmine as an example. Based on the multi-spectral remote sensing image of UAV (Unmanned Aer- ial Vehicle),the traditional VI (Vegetation Index) was extended by adding the red edge band,and the new vegetation index was put forward. Then,based on the corresponding VI and combind ynchronous field measurements of chloro- phyll data,nivariate and multivariate chlorophyll inversion models were constructed using empirical modelling meth- ods,respectively,the remote sensing inversion models of chlorophyll in corn leaves were built,validated,and com- pared. Finally,the spatial distribution of chlorophyll in corn leaves was analyzed using the best model in the study are- a. The results showed that the correlation between the vegetation index and the leaf chlorophyll after introducing the red edge was significantly enhanced compared to the traditional vegetation index. In all univariate models,MVIred-edge (Red-edge Modified Vegetation Index) which were added red edge were selected as the best vegetation index. Using nivariate and MLR ( Multivariate Stepwise Linear Regression) method, the precision of the models based on the MVIred-edge was improved compared to the VI with the calibration coefficient of determination (R2 ) raised between 0. 10 and 0. 20,and the calibration RMSE (Root Mean Squares Error) reduced between 0. 11 and 1. 98,the validation R2 raised between 0. 03 and 0. 12,the validation RMSE reduced between 1. 0 and 1. 5,and finally the EA ( Estimation Accuracy) reached 83. 4% . The MLR models composed of MVIred-edge ,GNDVI (Green Normalized Difference Vegeta- tion Index),NDVI(Normalized Difference Vegetation Index) obtained generally good performance with the EA greater than 80% . The main reasons improved the model precision were that red edge band are more ensitive to egetation hlo- rophyll esponses. Among them,when the vegetation biomass is large,the pigment content is high,and the growth is vig- orous,the red edge will move to the long wave direction (red shift). When the vegetation suffers from diseases,pests, pollution,leaf aging and other factors,the red edge will move to the short wave direction ( blue shift). It is indicated that the red-edge band is sensitive to the chlorophyll response of the crops in the coal mining subsidence area. The red edge band is introduced to construct the chlorophyll inversion model,which can be applied to remote sensing inversion of chlorophyll drones in maize leaves under the influence of coal mining subsidence. The study shows that the corn chlorophyll content in the coal mining subsidence basin is mainly concentrated in the 52 ~ 61 SPAD ( Soil and Plant Analyzer Development),and the area accounts for 75. 1% of the total. The chlorophyll content in the 9 ~ 52 SPAD crop area accounts for 5. 7% . The overall growth of the corn is affected by the disturbance of coal mining. The chloro- phyll content gradually decreased from the edge of the subsidence basin to the center. The study provides basic data and theoretical support for land damage monitoring and evaluation,land reclamation and ecological restoration.