During the last decades, researchers have developed novel computing methods to help viticulturists solve their problems, primarily those linked to yield estimation of their crops. This article aims to summarize the existing research associated with computer vision and viticulture. It focuses on approaches that use RGB images directly obtained from parcels, ranging from classic image analysis methods to Machine Learning, including novel Deep Learning techniques. We intend to produce a complete analysis accessible to everyone, including non-specialized readers, to discuss the recent progress of artificial intelligence (AI) in viticulture. To this purpose, we present work focusing on detecting grapevine flowers, grapes, and berries in the first sections of this article. In the last sections, we present different methods for yield estimation and the problems that arise with this task.

Plant-growth-promoting rhizobacteria (PGPR) could potentially enhance photosynthesis and benefit plant growth by improving soil nutrient uptake and affecting plant hormone balance. Several recent studies have unveiled a correlation between alterations in photosynthesis and host plant resistance levels. Photosynthesis provides materials and energy for plant growth and immune defense and affects defense-related signaling pathways. Photosynthetic organelles, which could be strengthened by PGPR inoculation, are key centers for defense signal biosynthesis and transmission. Although endophytic PGPRs metabolize plant photosynthates, they can increase soluble sugar levels and alternate sugar type and distribution. Soluble sugars clearly support plant growth and can act as secondary messengers under stressed conditions. Overall, carbohydrate metabolism modifications induced by PGPR may also play a key role in improving plant resistance. We provide a concise overview of current knowledge regarding PGPR-induced modulation in carbohydrate metabolism under both pathogen-infected and pathogen-free conditions. We highlight PGPR application as a cost-saving strategy amidst unpredictable pathogen pressures.

In many vineyards around the world, Botrytis cinerea (B. cinerea) causes one of the most serious diseases of aerial grapevine (Vitis vinifera L.) organs. The control of the disease relies mainly on the use of chemical products whose use is increasingly challenged. To develop new sustainable methods to better resist B. cinerea, beneficial bacteria were isolated from vineyard soil. Once screened based on their antimicrobial effect through an in vivo test, two bacterial strains, S3 and S6, were able to restrict the development of the pathogen and significantly reduced the Botrytis-related necrosis. The photosynthesis analysis showed that the antagonistic strains also prevent grapevines from considerable irreversible PSII photo-inhibition four days after infection with B. cinerea. The 16S rRNA gene sequences of S3 exhibited 100% similarity to Bacillus velezensis, whereas S6 had 98.5% similarity to Enterobacter cloacae. On the other hand, the in silico analysis of the whole genome of isolated strains has revealed the presence of “biocontrol-related” genes supporting their plant growth and biocontrol activities. The study concludes that those bacteria could be potentially useful as a suitable biocontrol agent in harvested grapevine.

Recognized as the causal agent of net blotch,Drechslera teresis responsible for major losses of barley crop yield. The consequences of this leaf disease are due to the impact of the infection on the photosynthetic performance of barley leaves. To limit the symptoms of this ascomycete, the use of beneficial bacteria known as “Plant Growth Promoting Rhizobacteria” constitutes an innovative and environmentally friendly strategy. A bacterium named as strain B25 belonging to the genusBurkholderiashowed a strong antifungal activity againstD. teres. The bacterium was able to limit the development of the fungus by 95% in detached leaves of bacterized plants compared to the non-bacterized control. In this study, in-depth analyses of the photosynthetic performance of young barley leaves infected withD. teresand/or in the presence of the strain B25 were carried out both in and close to the necrotic area. In addition, gas exchange measurements were performed only near the necrotic area. Our results showed that the presence of the beneficial bacterium reduced the negative impact of the fungus on the photosynthetic performance and modified only the net carbon assimilation rate close to the necrotic area. Indeed, the presence of the strain B25 decreased the quantum yield of regulated non-photochemical energy loss in PSII noted as Y(NPQ) and allowed to maintain the values stable of maximum quantum yield of PSII photochemistry known as F_v/F_mand close to those of the control in the presence ofD. teres. To the best of our knowledge, these data constitute the first study focusing on the impact of net blotch fungus and a beneficial bacterium on photosynthesis and respiratory parameters in barley leaves.

Efficient computation in deep neural networks is crucial for real-time object detection. However, recent advancements primarily result from improved high-performing hardware rather than improving parameters and FLOP efficiency. This is especially evident in the latest YOLO architectures, where speed is prioritized over lightweight design. As a result, object detection models optimized for low-resource environments like microcontrollers have received less attention. For devices with limited computing power, existing solutions primarily rely on SSDLite or combinations of low-parameter classifiers, creating a noticeable gap between YOLO-like architectures and truly efficient lightweight detectors. This raises a key question: Can a model optimized for parameter and FLOP efficiency achieve accuracy levels comparable to mainstream YOLO models? To address this, we introduce two key contributions to object detection models using MSCOCO as a base validation set. First, we propose LeNeck, a general-purpose detection framework that maintains inference speed comparable to SSDLite while significantly improving accuracy and reducing parameter count. Second, we present LeYOLO, an efficient object detection model designed to enhance computational efficiency in YOLO-based architectures. LeYOLO effectively bridges the gap between SSDLite-based detectors and YOLO models, offering high accuracy in a model as compact as MobileNets. Both contributions are particularly well-suited for mobile, embedded, and ultra-low-power devices, including microcontrollers, where computational efficiency is critical. Code: https://github.com/LilianHollard/LeYOLO