College of Geoscience and Surveying Engineering, China University of Mining and Technology (Beijing), Beijing, 100083, China.
Department of Land Management, Zhejiang University, Hangzhou, 310058, China.
Environ Monit Assess. 2020 Jun 30;192(7):474. doi: 10.1007/s10661-020-08453-5.
In eastern China, coal mining has damaged a large amount of farmland, posing a great threat to food security. Backfilling with coal waste, fly ash, and sediments from rivers is an effective method to restore farmland. This study was conducted at the reclaimed area (RA) and the undisturbed area (UA) in Shandong Province, China. Soil and plant analyzer development (SPAD) of corn was selected as an indicator of crop growth. Multi-spectral data was obtained by the unmanned aerial vehicle equipped with a camera. By analyzing the correlation between SPAD and spectral bands, the common vegetation index is improved. Different regression methods were used to construct the SPAD inversion model. The distribution of corn SPAD was monitored to objectively evaluate reclamation technology. The results are as follows: (1) the vegetation index improved using the red-edge band has a higher correlation with SPAD, and the largest coefficient of determination (R) value is 0.779; (2) the optimum inversion model for both jointing stage (R = 0.676) and milky stage (R = 0.661) is the linear regression model; the optimum model for both tasseling stage (R = 0.809) and filling stage (R = 0.830) is the partial least squares regression model; (3) the SPAD inversion map of RA and UA obtained by the optimum model shows that the corn grown in RA is slightly better than in UA. This study realized the rapid and efficient monitoring of the reclamation effects based on multi-spectral imagery and verified the feasibility of backfilling reclamation with Yellow River sediment in coal mining subsidence areas.
在中国东部，煤炭开采破坏了大量农田，对粮食安全构成了巨大威胁。采用煤矸石、粉煤灰和河流沉积物进行回填是恢复农田的有效方法。本研究在中国山东省的复垦区（RA）和未扰动区（UA）进行。玉米的土壤和植物分析器（SPAD）被选为作物生长的指标。多光谱数据由配备相机的无人机获取。通过分析 SPAD 与光谱波段之间的相关性，改进了常用植被指数。使用不同的回归方法构建了 SPAD 反演模型。监测玉米 SPAD 的分布，客观评价复垦技术。结果如下：（1）使用红边波段改进后的植被指数与 SPAD 相关性更高，最大决定系数（R）值为 0.779；（2）拔节期（R=0.676）和乳熟期（R=0.661）的最佳反演模型均为线性回归模型；抽雄期（R=0.809）和灌浆期（R=0.830）的最佳模型为偏最小二乘回归模型；（3）最优模型得到的 RA 和 UA 的 SPAD 反演图表明，RA 中生长的玉米略好于 UA。本研究基于多光谱图像实现了复垦效果的快速高效监测，验证了采用黄河泥沙进行采煤沉陷区回填的可行性。
