We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Identification of Farmland Mulching Film Types Based on UAV Multispectral Images and Analysis of Recycling Application: A Case Study of Hetao Irrigation District
Original title: 基于无人机多光谱影像的农田覆膜类型识别及回收应用分析——以河套灌区为例
Summary
This isn't really a human health study — it's about using drone imagery to map plastic sheeting ("mulch") that farmers lay over fields to help crops grow. The bigger picture matters for you, though: leftover plastic mulch that isn't properly collected can break down into microplastics that contaminate soil, water, and eventually the food supply, so better tools for tracking and recovering this plastic could help reduce that contamination over time.
The spatial distribution of plastic-mulched farmland is fundamental for verifying mulch input, preventing residual film pollution, organizing mechanized recovery, and planning resource-oriented recycling facilities. To address fragmented field patterns, coexisting mulch colors, interference from post-irrigation water and greenhouse structures, and the unclear engineering adaptability of semantic segmentation models in the Hetao Irrigation District, this study used five-band UAV multispectral orthomosaics acquired from 2022 to 2024. A five-class classification system was established, including non-cultivated land, other mulched farmland, black narrow-mulched farmland, unmulched farmland, and white narrow-mulched farmland. Ten representative semantic segmentation models were evaluated at spatial resolutions of 0.08, 1.35, 2.7, and 5.4 m. To approximate operational use, only radiometrically corrected original five-band reflectance was used as model input, without additional spectral indices, texture features, or external environmental variables.