Soil salinity estimation incorporating environmental covariables using UAV remote sensing for precision field management

Authors: MA, WEITONG - Northwest A&f University; HAN, WENTING - Northwest A&f University; CUI, XIN - Northwest A&f University; Zhang, Huihui; ZHANG, LIYUAN - Northwest A&f University; DONG, YUXIN - Northwest A&f University; ZHAI, XUEDONG - Northwest A&f University

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/8/2025
Publication Date: 5/22/2025
Citation: Ma, W., Han, W., Cui, X., Zhang, H., Zhang, L., Dong, Y., Zhai, X. 2025. Soil salinity estimation incorporating environmental covariables using UAV remote sensing for precision field management. Computers and Electronics in Agriculture. 237. Article e110532. https://doi.org/10.1016/j.compag.2025.110532.
DOI: https://doi.org/10.1016/j.compag.2025.110532

Interpretive Summary: Soil salinity is a significant problem that affects agriculture, food supply, and the environment. To effectively manage it, we need to accurately measure the level of salt in the soil. Currently, we use satellite technology to do this on a large scale, but it has limitations. This study explored a new approach using an unmanned aerial system. Our findings showed that certain indicators, especially those related to blue and red spectral bands in the images, were closely linked to soil salt levels. We also found that roughness on the soil surface had a significant impact. An advanced computer model performed better than traditional methods in soil salinity prediction with higher accuracy. In summary, this study has introduced a more precise way to measure soil salinity, combining environmental factors and advanced modeling technology. It's a step forward in helping us manage soil salinity more effectively.

Technical Abstract: Soil salinization poses a significant challenge to agricultural production, food security, and the ecological environment. Timely and precise identification of the extent and intensity of soil salinization is crucial for effective prevention and treatment. Currently, satellite remote sensing is the main tool for large-scale soil salinity monitoring. However, limitations in spatiotemporal resolution and environmental complexity leave room for improving the accuracy of regional scale soil salt content (SSC) monitoring. This study explored the potential of environmental covariates and spectral variables derived from multispectral images for estimating SSC. Aerial and field campaigns were conducted in 18 study areas in October 2021 and April 2022 on bare farmland, simultaneously capturing ground truth data for SSC, soil water content (SWC), and soil surface roughness (SSR). The sensitivity of eight salinity indices (SIs), three spectral indices, and two environmental covariates to SSC was analyzed. The optimal parameter combination was set as input variables, and SSC estimation was performed using a linear regression model and three machine learning regression methods. Results showed that salinity indices related to blue and red bands exhibited a strong correlation with SSC, while environmental covariate SSR showed an indirect correlation. Among the machine learning algorithms tested, all outperformed the linear regression model in multi-parameter SSC estimation, with the SVR_SSC model demonstrating the highest accuracy (R2 = 0.77). This study introduced a comprehensive method for SSC estimation that integrates environmental covariates and provides a valuable reference for the rapid and accurate assessment of soil salinization.