Over the last few years, several Convolutional Neural Networks for object detection have been proposed, characterised by different accuracy and speed. In viticulture, yield estimation and prediction is used for efficient crop management, taking advantage of precision viticulture techniques. Convolutional Neural Networks for object detection represent an alternative methodology for grape yield estimation, which usually relies on manual harvesting of sample plants. In this paper, six versions of the You Only Look Once (YOLO) object detection algorithm (YOLOv3, YOLOv3-tiny, YOLOv4, YOLOv4-tiny, YOLOv5x, and YOLOv5s) were evaluated for real-time bunch detection and counting in grapes. White grape varieties were chosen for this study, as the identification of white berries on a leaf background is trickier than red berries. YOLO models were trained using a heterogeneous dataset populated by images retrieved from open datasets and acquired on the field in several illumination conditions, background, and growth stages. Results have shown that YOLOv5x and YOLOv4 achieved an F1-score of 0.76 and 0.77, respectively, with a detection speed of 31 and 32 FPS. Differently, YOLO5s and YOLOv4-tiny achieved an F1-score of 0.76 and 0.69, respectively, with a detection speed of 61 and 196 FPS. The final YOLOv5x model for bunch number, obtained considering bunch occlusion, was able to estimate the number of bunches per plant with an average error of 13.3% per vine. The best combination of accuracy and speed was achieved by YOLOv4-tiny, which should be considered for real-time grape yield estimation, while YOLOv3 was affected by a False Positive–False Negative compensation, which decreased the RMSE.

The processes of soil erosion and land degradation are more rapid in the case of inappropriate agricultural management, which leads to increased soil loss rates. Moreover, climatic conditions are one of the most important determining factors affecting agriculture, especially in the Mediterranean areas featuring irregular rainfall and high summer temperatures. Conservation agriculture (CA) can make a significant contribution to reducing soil erosion risk on the annual cropland (ACL) of the Mediterranean region in comparison with conventional management (CM). The objective of this study is to provide soil loss rate maps and calculate the values for each altitude and slope class and their combination for the Apulia region in four annual production cycles for the scenarios CM and CA. The present study estimates the significance of the adoption of CA on soil erosion assessment at regional scale based on the Revised Universal Soil Loss Equation (RUSLE) model. The parameters of the RUSLE model were estimated by using remote sensing (RS) data. The erosion probability zones were determined through a Geographic Information System (GIS) and Google Earth Engine (GEE) approach. Digital terrain model (DTM) at 8 m, ACL maps of the Apulia region, and rainfall and soil data were used as an input to identify the most erosion-prone areas. Our results show a 7.5% average decrease of soil loss rate during the first period of adoption of the four-year crop cycle—from 2.3 t ha⁻¹ y⁻¹ with CM to 2.1 t ha⁻¹ y⁻¹ with the CA system. CA reduced soil loss rate compared to CM in all classes, from 10.1% in hill class to 14.1% for hill + low slope class. These results can therefore assist in the implementation of effective soil management systems and conservation practices to reduce soil erosion risk in the Apulia region and in the Mediterranean basin more generally.