Soil Degradation Mapping in Drylands Using Unmanned Aerial Vehicle (UAV) Data

Abstract

Arid and semi-arid landscapes often show a patchwork of bare and vegetated spaces. Their heterogeneous patterns can be of natural origin, but may also indicate soil degradation. This study investigates the use of unmanned aerial vehicle (UAV) imagery to identify the degradation status of soils, based on the hypothesis that vegetation cover can be used as a proxy for estimating the soils’ health status. To assess the quality of the UAV-derived products, we compare a conventional field-derived map (FM) with two modelled maps based on (i) vegetation cover (RGB map), and (ii) vegetation cover, topographic information, and a flow accumulation analysis (RGB+DEM map). All methods were able to identify areas of soil degradation but differed in the extent of classified soil degradation, with the RGB map classifying the least amount as degraded. The RGB+DEM map classified 12% more as degraded than the FM, due to the wider perspective of the UAV compared to conventional field mapping. Overall, conventional UAVs provide a valuable tool for soil mapping in heterogeneous landscapes where manual field sampling is very time consuming. Additionally, the UAVs’ planform view from a bird’s-eye perspective can overcome the limited view from the surveyors’ (ground-based) vantage point.

Agronomic Practices for Reducing Soil Erosion in Hillside Vineyards under Atlantic Climatic Conditions (Galicia, Spain)

“…Future work, including photogrammetric approaches using images captured by drones [44,45], could enhance the estimations obtained in the current study. These technologies provided successful results for mapping soil degradation in drylands, achieving a similar accuracy to conventional field mapping [44]. Nevertheless, geo-referencing and the geometric correction of images is imperative, requiring time, effort, a digital elevation model, and excellent ground control.…”

“…The use of grid control points distributed over the study site and correctly geo-referenced might be a reasonable solution for achieving enough accuracy for employing drones in the estimation of soil erosion, as proven for monitoring gullies in Morocco [45]. However, current studies employing these technologies referred to relatively large erosion processes (gullies) in arid environments [44,45], which might not occur in vineyards in which vegetation might pose problems for analyzing the images acquired. Therefore, the use of drone-acquired images for assessing erosion processes in vineyards still remains an interrogation.…”

Reliability of using high-resolution aerial photography (red, green and blue bands) for detecting available soil water in agricultural land

“…The KED method can map soil water content with very few samples at the high spatial resolution, using the best geophysical information. To improve the quality of soil water content mapping results, one can combine orthophoto with small-scale topographic information derived from DEM generated from UAV data (Krenz et al, 2019). Besides the direct use of these maps as variables for biophysical modeling, the uncertainty maps are an exciting source of information that can be used to define future soil sampling schemas (Padarian et al, 2017).…”

Facing Erosion Identification in Railway Lines Using Pixel-Wise Deep-Based Approaches

“…However, as introduced, performing such task using manual efforts is costly and automatic methods appear as an attractive alternative. In fact, over the years, several techniques [3,[14][15][16][17][18][19][20][21][22][23] have been proposed to perform erosion identification using remote sensing datasets.…”