Extraction of the upright maize straw by integrating UAV multispectral and DSM data

Upright maize straw left in the field during autumn and winter significantly contributes to severe air pollution in agricultural ecosystems due to burning. It is essential to obtain the spatial distribution of upright maize straw quickly and accurately for effective management and environmental prot...

Full description

Saved in:
Bibliographic Details
Main Authors: Aosheng Chao, Enguang Xing, Yunbing Gao, Cunjun Li, Yuan Qin, Chengyang Zhu, Yu Liu, Qingwei Zhu
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:International Journal of Applied Earth Observations and Geoinformation
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1569843225002699
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Upright maize straw left in the field during autumn and winter significantly contributes to severe air pollution in agricultural ecosystems due to burning. It is essential to obtain the spatial distribution of upright maize straw quickly and accurately for effective management and environmental protection. However, identifying upright maize straw using remote sensing is difficult because its spectral properties resemble those of other land covers like straw residue, bare soil, and sparse wheat at the same period. This study proposes a novel index for extracting upright maize straw by integrating low-cost unmanned aerial vehicle (UAV) visible to near-infrared spectral bands with digital surface model (DSM) data. First, we analyzed the spectral characteristics of four land cover types: upright maize straw, straw residue, bare soil, and sparse wheat, and proposed the adjusted straw index (ASI) that leverages green, red, and red-edge bands. Next, we combined DSM data with the ASI to develop the adjusted height straw index (AHSI), considering the height of the upright maize straw. Finally, the combination of index-plus-Otsu threshold segmentation and random forest (RF) methods was applied to identify and extract the spatial distribution of upright maize straw. The results showed that our method effectively detected the main regions of upright maize straw. The two proposed straw indices achieved over 87%(ASI) and 96%(AHSI) extraction accuracies across three different study regions. The two new indices not only significantly improve the accuracy of upright maize straw identification but also provide a new approach for low-cost UAV-based identification of non-photosynthetic vegetation (NPV).
ISSN:1569-8432