| Abstrakt: |
The fast and low-cost identification and determination of boundaries of subsided cultivated land in high-groundwater-level coal mines is an urgent need for land reclamation of coal mines. Accurate extraction of subsided cultivated land can provide basic data for boundary determination. With the use of unmanned aerial vehicle remote sensing image as the data source, the spectral characteristic values of typical ground objects in red, green, and blue bands were analyzed, and 10 visible-band vegetation indexes were selected and used to extract the cultivated land. Through vegetation index calculation and analysis, excess green index (EXG), vegetation index (VEG), combination index 2 (COM2), visible-band difference vegetation index (VDVI), red green blue vegetation index (RGBVI), and color index of vegetation (CIVE) were selected for information extraction of cultivated land. The following results were obtained: by visual judgment, VEG, EXG and CIVE were first excluded. Results obtained by using COM2, VDVI, and RGBVI indexes were relatively similar. The regions of interest of cultivated land were used to evaluate the accuracy, and the results were as follows, in decreasing order of accuracy: COM2 (85.27%), VDVI (81.05%), and RGBVI (71.38%). Another area was selected for the verification of the extraction by using COM2, and the results showed the cultivated land extraction accuracy is 91.90%, which indicated applicability for information extraction of subsided cultivated land in high-groundwater-level coal mines. This study provides basic data and reference method for the determination of boundaries and land reclamation of subsided cultivated land in high-groundwater-level coal mines. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink