AI predicts harvest times for farmers
A technological shift in agriculture: AI assists growers in optimizing harvest timing by forecasting ideal periods and factoring in weather conditions.
New artificial intelligence (AI) technology is providing farmers with tools to more accurately time their harvests. These models predict optimal picking periods by considering variable weather conditions, which is crucial for maximizing crop quality and revenue.
Joel Carter from Okanagan Specialty Fruits explains that traditional methods are insufficient. Last year, for example, the apple harvest in Washington State had to be halted due to extreme heat. "You need to know more than just when your fruit is ripe. How long do you have to pick it? These models are really helpful," Carter stated. The company is already experimenting with cameras from Canadian firm Vivid Machines, which capture imagery of apple trees for AI analysis.
While AI models offer more precise forecasts, their accuracy depends on the quality of the historical data fed into the system. Carter emphasizes that the system must be customized for each farm. Apples have a harvest window of about three weeks, but for berries like strawberries, the timeframe can be as short as a few days.
UK-based FruitCast offers similar forecasts for strawberry, raspberry, blackberry, blueberry, and tomato harvests. CEO Raymond Martin says their system analyzes footage from drones, smartphones, or cameras mounted on farm vehicles. "We do exactly what the farmers could do, but on a scale they cannot," Martin explained.
FruitCast's models incorporate weather and irrigation conditions into their predictions. The company claims its forecasts are 90% accurate one week out and 83% accurate three weeks out regarding actual harvested volume. Despite rapid technological advancements, industry experts like Neill Finlayson of Angus Soft Fruits note that this is still an ongoing journey, and a fully integrated forecasting ecosystem has yet to be achieved.