Soil is a crucial ecosystem component, supporting plant growth and contributing to environmental quality. Digital maps of soil properties play a critical role in effective soil conservation planning and management. This paper comprehensively reviews recent developments in Digital Soil Mapping (DSM) models from 2017 to 2022, focusing on sustainable agriculture and environmental management. We outline the DSM process, starting with data acquisition and explaining the soil-forming factors and the SCORPAN model. We examine different data sources used in DSM and their applications. We then discuss the methods for mapping soil, including geostatistical, machine learning, and deep learning, highlighting their strengths and limitations. We also explore the introduction of transfer learning in DSM and its potential for enhancing accuracy. Additionally, we review how recent studies validate and estimate uncertainty in their results. To analyze current trends in DSM, we perform statistical analysis on the reviewed works. Finally, we compare the findings of several exploratory studies and identify remaining challenges and future opportunities in DSM research. Overall, this review provides valuable insights into recent developments in DSM and their potential applications in soil management and conservation.
Belkadi, W., Drias, Y. (2023). Advancements in Digital Soil Mapping: From Data Acquisition to Uncertainty Estimation - A Comprehensive Review. In Artificial Intelligence Doctoral Symposium 5th Doctoral Symposium, AID 2022, Algiers, Algeria, September 18–19, 2022, Revised Selected Papers (pp.162-177). Springer Science and Business Media Deutschland GmbH [10.1007/978-981-99-4484-2_13].
Advancements in Digital Soil Mapping: From Data Acquisition to Uncertainty Estimation - A Comprehensive Review
Drias Y.
2023
Abstract
Soil is a crucial ecosystem component, supporting plant growth and contributing to environmental quality. Digital maps of soil properties play a critical role in effective soil conservation planning and management. This paper comprehensively reviews recent developments in Digital Soil Mapping (DSM) models from 2017 to 2022, focusing on sustainable agriculture and environmental management. We outline the DSM process, starting with data acquisition and explaining the soil-forming factors and the SCORPAN model. We examine different data sources used in DSM and their applications. We then discuss the methods for mapping soil, including geostatistical, machine learning, and deep learning, highlighting their strengths and limitations. We also explore the introduction of transfer learning in DSM and its potential for enhancing accuracy. Additionally, we review how recent studies validate and estimate uncertainty in their results. To analyze current trends in DSM, we perform statistical analysis on the reviewed works. Finally, we compare the findings of several exploratory studies and identify remaining challenges and future opportunities in DSM research. Overall, this review provides valuable insights into recent developments in DSM and their potential applications in soil management and conservation.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
