利用高光谱反演、 监测植被生长状况的基础是光谱特征识别。 以半干旱采煤塌陷区为样地, 利用Field Spec 3地物光谱仪与SPAD-502叶绿素仪同步采集采煤地表塌陷形成的不同应力区(非采区、 中性区、 拉伸区、 压缩区)典型植物叶片光谱反射率与叶绿素含量(SPAD值), 分析典型植物相同应力区SPAD值升高其光谱特征的变化, 对比不同应力区典型植物SPAD值较高与较低时光谱特征的差异, 并借助Matlab软件深入研究不同应力区典型植物SPAD值与差值指数、 归一化指数的相关关系。 结果表明: (1)不同应力区同种植物光谱曲线随着SPAD值不同变化规律相异, 可见光波段区分明显, 其余波段受应力影响的区域样本SPAD值不同, 反射率比非采区波动更为剧烈与无序。 可见光波段, 糙隐子草、 柠条、 杨树、 油蒿SPAD值低的样本光谱曲线绿峰缺失, SPAD值升高, 绿峰出现但位置红移, SPAD值高于30时, 为典型植被光谱曲线, 油松样本SPAD值越高反射峰值越小; 受应力影响的区域SPAD值低的样本谷、 峰、 边特征参数缺失更多, 光谱变化规律不强。 (2)400~700 nm波段, 不同应力区糙隐子草、 油蒿、 油松、 柠条样本SPAD值较低组反射率显著高于较高组, 杨树样本相反; 780~1 350 nm波段, 拉伸区的糙隐子草、 非采区的油蒿和柠条、 压缩区的油松和杨树样本SPAD值较高组与较低组的反射率差异小; 相较于非采区, 受应力影响的糙隐子草、 油蒿、 柠条样本在所测波段SPAD值较高组与较低组的同波段反射率差值显著减小。 (3)受应力影响的区域样本SPAD值与光谱指数相关性较之非采区在某些波段大面积增强。 与非采区相比, 中性区的油蒿、 油松、 柠条、 杨树样本SPAD值与光谱指数的最大相关系数值均增大, 糙隐子草相反; 非采区植物SPAD值与NDVI最大相关系数均高于DI, 波段组合多位于近红外, 受应力影响区域的样本最大相关系数多数位于可见光波段。 本研究为矿区不同应力区典型植物高光谱波段识别与植物健康状态监测、 矿区生态环境精准治理提供了理论支撑。

The use of hyperspectral technology to invert and monitor vegetation is based on the identification of its spectral characteristics. The spectral reflectance and the SPAD values of leaves of typical plants, collected in different mining disturbance zones of Coal mining subsidence areas in semi-arid regiens, were simultaneous measured by Field Spec 3 spectrometer and the SPAD-502 chlorophyll meter to explore the spectral characteristics of the plants in the same zone with the SPAD values changing and the spectral differences of typical plants in different mining disturbance zones when the SPAD values was in the same interval. Furthermore, the correlations between SPAD values of leaves of typical plants and the spectral index were analyzed by using Matlab software. The results showed that the spectral characteristics of the same plant species in different mining disturbance zones were distinct with the SPAD values changing, and the spectral curves of different sample zones were clearly distinguished in visible band. In visible band, the green peak of spectral curves of Cleistogenes squarrosa, Artemisia ordosica, Caragana microphylla and Populus tremula in the non-mining zone were absent when the SPAD values was very low, and it appeared with the increase of SPAD value, but the position was red-shifted, when the SPAD value was above 30, the spectral features of the blue valley, green peak, red valley and red edge were obvious, and the higher the SPAD value of leaves of Chinese pine, the lower the spectral reflectance, while no rule was observed in the other zones. In addition, the spectral reflectance of samples with lower SPAD values of Cleistogenes squarrosa, Artemisia ordosica, Chinese pine and Caragana microphylla in the different mining disturbance zones was significantly higher than that of the samples with higher SPAD values from 400 to 700 nm band in general, but the trend of leaves of Populus tremula was just the opposite. The reflectance of the higher and lower groups of SPAD values of Cleistogenes squarrosa in the drawing zone, Artemisia ordosica and Caragana microphylla in the non-mining zone, and Chinese pine and Populus tremula in the compression zone were similar. Compared with the non-mining zone, the differences between the reflectance of the higher and lower groups of SPAD values of the Cleistogenes squarrosa, Artemisia ordosica, and Caragana microphylla in the mining disturbance zones significantly reduced. Moreover, the correlation between the SPAD value and the spectral index of the samples in the mining disturbance zones was enhanced in some bands than that in non-mining zone. The maximum correlation coefficient values between the SPAD value and the spectral index of Artemisia ordosica, Chinese pine, Caragana microphylla and Populus tremula had become lager in neutral zone than that in the non-mining zone, while that of Cleistogenes squarrosa was just opposite. At the same time, the maximum correlation coefficient between SPAD value and NDVI in non-mining zone was higher than that of DI, and the band combinations were mostly in the near infrared, but in the other zones, the band combinations were mostly in visible band. The results of this study provided theoretical support for identifying the differences in the spectral characteristics of typical plants in different mining disturbance zones, further monitoring the health status of plants and accurately controlling the ecological environment in mining area.