遥感
作物
地理
索引(排版)
地图学
环境科学
自然地理学
林业
计算机科学
万维网
作者
Xianwei Ma,Yanmin Shuai,Latipa Tuerhanjiang
标识
DOI:10.1080/17538947.2025.2538818
摘要
Plastic mulch (PM) is extensively utilized in agriculture to significantly increase crop yields. Nevertheless, residual mulch induces potentail risks to food security and environmental health. Therefore, precise identification of PM farmland (PMF) is essential for effective PM recycling and agricultural management. Remote sensing-based PM indices are efficacious for large-scale PMF identification, though the accuracy is impeded by complex cropping backgrounds and diverse PM types. By enhancing PM information while suppressing non-PM background effects, we proposed an adaptive plastic-mulched crop field index (APMI) for identifying black and white mulched field (BPMF/WPMF) from spectral prior knowledge of different PM types and their separability in the visible, infrared, and shortwave infrared bands with Landsat-8 OLI data. We selected four representative BPMF or WPMF areas with diverse terrains and crops to verify performance. The results indicate APMI can concurrently identify both BPMF and WPMF, and outperforms existing indices. In homogenous planting regions, overall accuracy (OA) reaches 98.02% and 96.04% (F1 scores = 96.43%/94.31%). In more complex backgrounds, OA remains above 94.44% and 86.91% (F1 scores = 92.60%/85.88%). Consequently, APMI demonstrates robust applicability in identifying various PMF and the significant potential for precise PMF mapping with medium-resolution imagery.
科研通智能强力驱动
Strongly Powered by AbleSci AI