Please use this identifier to cite or link to this item:
https://elib.bsu.by/handle/123456789/329022| Title: | Цифровое почвенное картографирование по данным воздушного лазерного сканирования |
| Other Titles: | Digital soil mapping using air laser scanning data |
| Authors: | Сазонов, Алексей Александрович Киндеев, Аркадий Леонидович Князев, Иван Сергеевич |
| Keywords: | ЭБ БГУ::ЕСТЕСТВЕННЫЕ И ТОЧНЫЕ НАУКИ::География ЭБ БГУ::ОБЩЕСТВЕННЫЕ НАУКИ::Информатика |
| Issue Date: | 2025 |
| Publisher: | Томский государственный университет |
| Citation: | Сазонов, А.А. Цифровое почвенное картографирование по данным воздушного лазерного сканирования / А.А. Сазонов, А.Л. Киндеев, И.С. Князев // Геосферные исследования. - 2025. - № 1. - С. 124–135. |
| Abstract: | Предложена методика прогнозного цифрового картографирования почвенного покрова на основе данных воздушного лазерного сканирования и мультиспектрального космического снимка. Методика была апробирована на территории Воложинского района Минской области Республики Беларусь. Для прогнозирования почвенного покрова была использована модель машинного обучения Random Forest. В результате исследования были получены две цифровые почвенные карты (по гидроморфизму и по типам почв). |
| Abstract (in another language): | Soil mapping is a critical task for a variety of applications, including agriculture, ecology, and environmental management. Traditionally, soil mapping has been conducted using field surveys, which are time-consuming, labor-intensive, and expensive. In recent years, there has been a growing interest in the use of remote sensing data for digital soil mapping (DSM). DSM methods use data from satellites, aircraft, or drones to generate digital soil maps with high spatial resolution and low cost. Methods: This study proposes a predictive DSM methodology based on airborne laser scanning (ALS) and multispectral satellite imagery. The study was conducted in the Volozhinsky District of the Minsk Region of Belarus. ALS data were used to derive morphometric indicators of relief, while multispectral satellite imagery were used to derive vegetation indices. A Random Forest machine-learning model was used to predict soil types and hydromorphic classes. Results: The results of the study showed that the proposed methodology was able to generate accurate digital soil maps with high spatial resolution. The overall accuracy of the soil type map was 64%, and the overall accuracy of the hydromorphic class map was 60 %. Discussion: The results of this study confirm the effectiveness of the proposed predictive DSM methodology. The methodology is based on a combination of ALS and multispectral satellite imagery, which provides a comprehensive set of data for soil mapping. The use of a Random Forest machine-learning model also contributes to the accuracy of the methodology. Conclusion: The proposed methodology is a promising approach for DSM in a variety of settings. The methodology is relatively easy to implement and can be used to generate digital soil maps at a variety of spatial resolutions. |
| URI: | https://elib.bsu.by/handle/123456789/329022 |
| Licence: | info:eu-repo/semantics/openAccess |
| Appears in Collections: | Кафедра почвоведения и геоинформационных систем (статьи) |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.