Cornell Engineering Students Tackle Vineyard Data Needs
By Tim Martinson
Two Cornell Engineering graduate students are using their engineering skills to develop prototype products for grape producers.
Hunter Adams (PhD Aerospace Engineering) and Jonathan Jaramillo (PhD Electrical and Computer Engineering) demonstrated equipment designed to assist growers in making management decisions in their vineyards at the recent Digital Viticulture: New Tools for Vineyard Management field day, held on July 17th at Anthony Road Vineyards on Seneca Lake near Penn Yan, NY.
On-site distributed weather stations.

Access to reliable weather data is crucial to grape growers for needs as diverse as frost and winter injury monitoring and disease forecasting. Currently, growers rely on weather stations that are often networked through the NEWA system. But their cost (around $2000) limits the number and placement of these stations to a few strategic locations. What if they were small and cheap enough so that growers could deploy several units at each farm or vineyard block?
Hunter Adams has developed a weather station called Monarch that fits in your wallet, and that costs about a twentieth of the price of the NEWA weather stations. At this size and price point, vineyards will be able to deploy tens of weather stations instead of a single weather station, and place them inside the grape canopy. This will allow growers to measure within-vineyard variations in temperature, relative humidity, and wetness – allowing more accurate tracking of frost and mid-winter low temperatures and disease development.
After finishing his degree in December, Hunter will be pursuing commercialization of the Monarch stations full-time, and is actively seeking out potential customers and grant support to fund continuing development of the sensors. His project has been funded through the National Science Foundation’s NSF Innovation Corps program, grant #1643287.
Low-cost crop estimation with cell phone imagery.

Obtaining accurate crop estimates has been a thorny problem for grape growers. Yield potential varies greatly from year to year – and also within vineyard blocks. Growers currently estimate their crop by manually counting grape clusters on a small number of vines – a time-consuming and labor-intensive process. What if they could mount a cell phone on an ATV or four-wheeler and use images to estimate the number of clusters on a much larger sample of vines?
Jonathan Jaramillo, working with faculty programs in Viticulture (Justine Vanden Heuvel) and Electrical and Computer Engineering (Kirstin Petersen) is using image analysis of cell phone video files and machine learning to identify and count grape clusters early in their development.
Current crop estimation practices can over or underestimate the final crop by 20-30% in some years – manual counting is time consuming, and no one can sample enough vines to be sure their estimate is accurate. The goal of this project is to reduce that error down to 5% or less.
Jonathan’s efforts are part of a larger collaboration between the Vanden Heuvel and Petersen programs to apply ‘soft touch’ robotic sensors to examine and measure grape cluster growth and attributes later in the growing season. The project is supported by the Cornell Initiative for Digital Agriculture’s Research Innovation fund and NIFA award #2019-67021-29225.

Tim Martinson is a senior extension associate in the Section of Horticulture, based at the NYS Agricultural Experiment Station in Geneva, NY.