Precision Pruning: How AI and Augmented Reality are Reshaping Vineyard Management

In the meticulous and tradition-steeped world of viticulture, technological innovation holds the promise of radical transformation. A Swiss startup, 3D2cut, is on the brink of such a change with its development of augmented reality (AR) eyewear software. This cutting-edge tool is designed to guide pruners with instant and precise identification of the vine shoots that need to be cut for a single Guyot system while respecting the sap flow of the plant. While still under development, this technology could mark a new era in the precision and efficiency of one of the most critical tasks in vine cultivation.

Pruning, a process seemingly straightforward to the casual onlooker, harbors complexities that only the most seasoned viticulturists can adeptly navigate. Into this scenario enters ViNet, an artificial intelligence (AI)-driven software for pruning assistance through augmented reality, stirring significant interest within the industry, though it has yet to hit the market.

Philibert Frick, founder and chair of 3D2cut, with extensive experience in both the industrial-financial sector and viticulture, spearheads this project that patiently weaves technological innovation with the wisdom of viticultural tradition. Launched in 2018, the company has progressed with caution, securing its first specialist employee in 2020 and working closely with the Institute for Perceptive AI (IDIAP) to develop a model of the vine structure.

At the heart of this technology is the ViNet software, capable of analyzing a vine's image to distinguish structural elements — trunk, shoots, buds — and suggest optimal cuts through a system of computer vision and decision-making rules, based on vine type and cultivation method. This tool not only promises enhanced pruning efficiency and precision but also aims to democratize respectful pruning methodology, making it accessible to a broader range of growers.

Massimo Giudici, master pruner and partner at 3D2cut, vouches for the technology's potential. After field-testing ViNet, he highlighted the software's robustness, capable of identifying plant parts in milliseconds without the need for additional backdrops or markers. Yet, challenges remain, such as adapting to varying light and weather conditions and accurately identifying bud orientation — crucial for successful pruning.

Balancing human perception with the AI 'brain' is also of significance in this innovation. While humans can easily identify shoots, AI can apply a set of decision rules more consistently to propose cuts that optimally shape the plant. Thus, the AR glasses are envisioned as a guide for the pruner, who ultimately decides whether to follow the provided recommendations.

Looking ahead, 3D2cut aspires to not only provide a pruning aid but also to explore potential applications such as developing pruning robots or employing drones in viticulture. The company's presence at the upcoming Sitevi exhibition is an opportunity to forge industrial alliances to propel these innovations.

The advancement of AI in viticulture opens a realm of possibilities where technological precision and respect for the plant's natural processes can coexist, enabling pruning that is not only efficient but also contributes to sustainability and the enhancement of wine quality.