Agricultural terraces are important for agricultural production and soil-and-water conservation. They comprise treads and risers that require manual construction and maintenance. If managed improperly, risers will collapse, causing soil loss, gully erosion, and cultivation threats. However, mapping terrace risers remains a challenge. This study presents a novel approach to automatically map terrace risers by combining remote sensing images and digital elevation models (DEMs). First, a terraced hillslope was extracted via a hill-shading method and edges in the image were detected using a Canny edge detector. Next, the DEM was used to generate the contour direction, and edges along this direction were searched and coded as candidate terrace risers via directional detection. Finally, the results of directional detection and the edge image obtained from the Canny detector were overlaid to backtrack complete terrace risers. The approach was validated using four study areas with different topographic characteristics in the Loess Plateau, China. The results verify that the approach achieves outstanding performance and robustness in mapping terrace risers. The precision, recall, and F-measure were 90.81%–97.57%, 88.53%–94.10%, and 90.13%–95.80%, respectively. This approach is flexible and applicable with freely available images and DEM sources.

The gully shoulder line, one of the most important topographical features, reveals the erosional process and reflects the geomorphological evolution of a loess area. Existing shoulder-line extraction methods are based on local window-filter or image-edge detection, which are sensitive to image noises and algorithm parameters, causing unsatisfactory accuracy and efficiency. This paper proposes a bidirectional relief-shading (BRS) method for loess shoulder-line extraction based on a 5-m Digital Elevation Model (DEM). First, two hill-shaded images are simulated with appropriate altitude and symmetric direction of the light. Second, the grey value difference between inter-gully and inner-gully area can be easily identified by fusing the two images with the mean method. Finally, the shoulder line can be derived by image segmentation using a threshold determined by an empirical equation. Experiments in three areas of loess tableland in Shaanxi Province validated the method, which has the advantages of being a simple operation with relatively high accuracy, 89.7% compared with manual digitalization, and high efficiency. We discuss three parameters in this method: zenith angle, azimuth angle, and segment threshold. Results suggest that the method is applicable for broad-scale gully shoulder line extraction in loess tablelands.

Image collection strategies and ground control points (GCPs) are of particular importance for uncrewed aerial vehicle combined with Structure‐from‐Motion (UAV–SfM) photogrammetry, and the generalization of their effects has proved elusive. This study designed various photogrammetric scenarios to investigate the effects of image collection strategies, ground control quantity, and their interaction on digital elevation model (DEM) errors and their spatial structure in high‐relief terrain. The results of 1.77 × 10⁵ UAV–SfM scenarios provide insights for improving UAV–SfM practices. A high image capture angle (20–40°) enhances camera calibration quality decreasing the magnitude and spatial correlation of errors. High camera inclination reduces the sensitivity of mean and standard deviation of error to flying height but not the spatial correlation of error. Including additional data (e.g. supplemented convergent images; images captured at multiple flying heights) has only a minor effect if imagery is highly inclined. GCPs provide more effective constraints than image collection strategies. The mean error and standard error decline quickly with a small number of GCPs and then become stable in all scenarios, but the spatial correlation of error can be further improved with increasing GCPs. However, the effects of GCP quantity do interact with image collection strategies. High camera inclination reduces requirements for GCPs, whilst strategies combining different flying heights and image orientations have little effect on necessary GCP quantity. The distribution of GCPs still affects the errors, but the effect of GCP distribution becomes less important with an increase in the number of GCPs. Finally, we show that UAV–SfM photogrammetric quality assessment should routinely assess the spatial dependence of error using a statistic like Moran's I.

Accurate assessment of the extent of crop distribution and mapping different crop types are essential for monitoring and managing modern agriculture. Medium and high spatial resolution remote sensing (RS) for Earth observation and deep learning (DL) constitute one of the most major and effective tools for crop mapping. In this study, we used high-resolution Sentinel-2 imagery from Google Earth Engine (GEE) to map paddy rice and winter wheat in the Bengbu city of Anhui Province, China. We compared the performance of different popular DL backbone networks with the traditional machine learning (ML) methods, including HRNet, MobileNet, Xception, and Swin Transformer, within the improved DeepLabv3+ architecture, Segformer and random forest (RF). The results showed that the Segformer based on the combination of the Transformer architecture encoder and the lightweight multilayer perceptron (MLP) decoder achieved an overall accuracy (OA) value of 91.06%, a mean F1 Score (mF1) value of 89.26% and a mean Intersection over Union (mIoU) value of 80.70%. The Segformer outperformed other DL methods by combining the results of multiple evaluation metrics. Except for Swin Transformer, which was slightly lower than RF in OA, all DL methods significantly outperformed RF methods in accuracy for the main mapping objects, with mIoU improving by about 13.5~26%. The predicted images of paddy rice and winter wheat from the Segformer were characterized by high mapping accuracy, clear field edges, distinct detail features and a low false classification rate. Consequently, DL is an efficient option for fast and accurate mapping of paddy rice and winter wheat based on RS imagery.