农业工程学报2018，Vol.34Issue(1)：232-240,9.DOI:10.11975/j.issn.1002-6819.2018.01.032
基于多源时序NDVI的稀土矿区土地毁损与恢复过程分析
Analysis of land damage and recovery process in rare earth mining area based on multi-source sequential NDVI
摘要
Abstract
The problem of land and ecological environment damage caused by ion mining has attracted more and more attention. Taking the Lingbei rare earth mining area as an example, combined with remote sensing normalized difference vegetation index (NDVI) time sequence analysis method, we used HJ-1B CCD (charge-coupled device), Landsat 5 and Landsat 8 remote sensing data during 1990-2016, and analyzed land distribution and rare earth mining area damage and recovery process. In order to reduce the NDVI error caused by the sensor itself, NDVI conversion equation of HJ-1B CCD and Landsat 5/8 data was constructed by using the regression analysis method, NDVI image of HJ-1B CCD was transformed into the NDVI image corresponding to the years of Landsat 5/8 (the analog image) by using conversion equation, and the root mean square error (RMSE) between the simulated image and the real Landsat 5/8 NDVI image was calculated and the accuracy of the transformation equation was verified. The results showed that theR2 values of NDVI equations from HJ-1B CCD and Landsat 5, HJ-1B CCD and Landsat 8 data were 0.9319 and 0.9552, both more than 0.9, indicating that there is obvious linear positive correlation between the NDVI of HJ-1B CCD and Landsat 5/8.The RMSEvaluesof NDVI imaging between simulation images and real Landsat5,andsimulation images and real Landsat8 were 0.0293and0.040,whichwere less than 0.05, showing that the obtained conversion equation has high accuracy and can eliminate the influence of different data sources on the NDVI. Combining NDVI threshold of mining and non-mining, and threshold of non-rare earth mining interference and rare earth mining interference, Lingbei rare earth mining area had annual rare earth mining activity from 1990 to 2016, rare earth mining area of which was smaller in 1990, 1991,1995 and 2000, mining area was more than 1 km2 in 2001-2006, and the mining area reached 2.5461 km2 in 2006; after 2006, to maintain the rare earth mining scale at the lower level, the country began to implement control of rare earth mining, and the spatial distribution of rare earth mining was more dispersed, which not only caused difficulties in supervision and waste of resources, but also increased governance difficulties to a certain extent. Testing the extraction accuracy of land destruction and restoration type, extraction accuracy of various types was found above 85％, most of which reached more than 90％, satisfying the requirement of accuracy. Analysis of land destruction and restoration type showed that the area that was not disturbed was 97.0821 km2, accounting for 45.41％ of the total mining area; anthropogenic disturbance area was 116.7093 km2, accounting for 54.59％ of the total mining area; the area affected by deforestation disturbance was the largest, which was 86.5341 km2, and the average time length of vegetation restoration was 3 years; the arable land that was not abandoned occupied 5.4045 km2; cultivated land abandoned for many years and that for returning to forest occupied 9.3906 km2; orchard development area was 3.1482 km2, and the average time length of vegetation restoration for 7 years; rare earth mining occupied 11.3544 km2, whose average recovery time was 11 years, and the recovery of the reclamation area was 5.0049 km2, so there was still 6.3495 km2 area with the regional vegetation that did not be recovered, and relevant departments should pay attention to the situation.关键词
土地复垦/土地利用/遥感/多源时序NDVI/稀土矿区/转换方程/时序分析法Key words
land reclamation/land use/remote sensing/multi-source sequential NDVI/rare earth mining area/conversion model/time-series analysis分类
矿业与冶金引用本文复制引用
李恒凯,雷军,吴娇..基于多源时序NDVI的稀土矿区土地毁损与恢复过程分析[J].农业工程学报,2018,34(1):232-240,9.基金项目
国家自然科学基金(41561091) （41561091）
江西省自然科学基金(20161BAB206143) （20161BAB206143）
江西省社会科学规划课题(17YJ20) （17YJ20）
江西省高校人文社会科学研究项目(JC17111) （JC17111）
