Olive agro-ecosystems in southern Italy have been heavily damaged due to Xylella fastidiosa subsp. pauca (Xfp). Replacing the Xfp-infected olive-growing areas with economically viable fruit tree species is thought to be a practical control measure. A land suitability analysis can provide an appropriate evaluation of a crop’s suitability in these areas. We evaluate the suitability of almond (Prunus dulcis B.), fig (Ficus carica L.), hazelnut (Corylus avellana L.), kiwifruit (Actinidia chinensis P.), pistachio (Pistacia vera L.), and pomegranate (Punica granatum L.) as fruit tree species immune/resistant to Xfp to be planted within the Xfp-infected olive-growing areas in the Apulia region to compensate for economic and environmental losses. Climate and soil data were used to carry out the land suitability analysis. We combined information for each parameter to obtain the overall suitability maps for the six proposed fruit tree crops using GIS (Geographic Information System). The analysis showed that the Xfp-infected olive-growing areas are suitable for the plantation of most of the proposed fruit tree crops, with different suitability levels as the climate and soil conditions vary among the study areas. In particular, large olive-growing areas are suitable for the cultivation of pomegranate (268,886 ha), fig (103,975 ha), and almond (70,537 ha), followed by kiwifruit (43,018 ha) and pistachio (40,583 ha). Hazelnut, with just 2744 ha of suitable land, was the species with fewer suitable areas in these semi-arid environments. This is the first study to provide practical containment measures against the diffusion of Xfp in southern Italy. Our results can help in the selection of the right immune/resistant tree species for replanting in Xfp-infected zones, therefore providing guidelines within the decision-making process to encourage the planting of some underrepresented fruit tree crops with viable economic values as well.

Ground vehicles equipped with vision-based perception systems can provide a rich source of information for precision agriculture tasks in orchards, including fruit detection and counting, phenotyping, plant growth and health monitoring. This paper presents a semi-supervised deep learning framework for automatic pomegranate detection using a farmer robot equipped with a consumer-grade camera. In contrast to standard deep-learning methods that require time-consuming and labor-intensive image labeling, the proposed system relies on a novel multi-stage transfer learning approach, whereby a pre-trained network is fine-tuned for the target task using images of fruits in controlled conditions, and then it is progressively extended to more complex scenarios towards accurate and efficient segmentation of field images. Results of experimental tests, performed in a commercial pomegranate orchard in southern Italy, are presented using the DeepLabv3+ (Resnet18) architecture, and they are compared with those that were obtained based on conventional manual image annotation. The proposed framework allows for accurate segmentation results, achieving an F1-score of 86.42% and IoU of 97.94%, while relieving the burden of manual labeling.