RESUMEN
Even though mechanization has dramatically decreased labor requirements, vineyard management costs are still affected by selective operations such as winter pruning. Robotic solutions are becoming more common in agriculture, however, few studies have focused on grapevines. This work aims at fine-tuning and testing two different deep neural networks for: (i) detecting pruning regions (PRs), and (ii) performing organ segmentation of spur-pruned dormant grapevines. The Faster R-CNN network was fine-tuned using 1215 RGB images collected in different vineyards and annotated through bounding boxes. The network was tested on 232 RGB images, PRs were categorized by wood type (W), orientation (Or) and visibility (V), and performance metrics were calculated. PR detection was dramatically affected by visibility. Highest detection was associated with visible intermediate complex spurs in Merlot (0.97), while most represented coplanar simple spurs allowed a 74% detection rate. The Mask R-CNN network was trained for grapevine organs (GOs) segmentation by using 119 RGB images annotated by distinguishing 5 classes (cordon, arm, spur, cane and node). The network was tested on 60 RGB images of light pruned (LP), shoot-thinned (ST) and unthinned control (C) grapevines. Nodes were the best segmented GOs (0.88) and general recall was higher for ST (0.85) compared to C (0.80) confirming the role of canopy management in improving performances of hi-tech solutions based on artificial intelligence. The two fine-tuned and tested networks are part of a larger control framework that is under development for autonomous winter pruning of grapevines. Supplementary Information: The online version contains supplementary material available at 10.1007/s11119-023-10006-y.
RESUMEN
Lonicera japonica is a ubiquitous medicinal species in China.Winter pruning has long been used to improve its quality and yield, but the mechanism is rarely studied.Therefore, in this study, the growth phenotypes of L.japonica processed with different pruning methods were observed and the yield-and quality-boosting mechanism of pruning was analyzed.Specifically, the young shoots of the three-year old L.japonica were cut to different degrees(heavy pruning, mild pruning, and no pruning, respectively) in winter in 2020 and 2021, respectively, and the growth phenotypes, hormone content, and gene expression of the lateral buds at the sprouting stage and young shoots at the anthesis stage in the next year were analyzed.The result showed that the length, flower bud number, internode length, and node number of young shoots in the next year were in the order of heavy pruning>mild pruning>no pruning.The content of auxin and zeatin in apical buds of young shoots at the anthesis stage was the highest in the heavy pruning group, followed by the mild pruning group, and coming in the third was the no pruning group.The content of auxin and zeatin in lateral buds at the sprouting stage was in the order of no pruning>mild pruning>heavy pruning.Transcriptome analysis of the lateral buds at sprouting stage yielded the differentially expressed genes related to auxin and cytokinin, such as Lj1A1163T36, Lj3A719T115, Lj7C657T7, Lj9C505T15, and Lj9A505T70.In conclusion, the growth phenotypes of young shoots of L.japonica processed with different pruning methods in winter were related to the difference in hormone content in the apical buds.Therefore, winter pruning influenced the content of auxin and cytokinin in new shoots of L.japonica and further regulated the expression of hormone-related genes, thereby promoting shoot growth and increasing the yield of L.japonica.