2016
FruitSpec’s solution for highly accurate estimates/projections of fruit yields is based on hyperspectral machine vision technology and deep learning algorithms. Specially-designed FruitSpec sensor pods are mounted on both sides of the tractor and scan the trees as the tractor moves along the orchard rows during regular activity. Applied computer vision and an AI algorithm automatically count and estimate fruit number and size. The customer (grower, packing house) receives reports regarding precise fruit yield and size distribution at both the individual tree level (lowest resolution) and the overall area scanned (highest resolution). FruitSpec allows better decision making using accurate data. Right from the beginning of the season, logistics, marketing and agronomy management decisions can be based on accurate yield estimates .
Today, key business decisions along the fruit value chain – from growers to retailers – are based on yield estimations performed by farmers/workers using a visual “count” from a sampling of a few trees. These estimations are inaccurate and often result in damaging business decisions. With dramatic fluctuations in fruit yield between seasons (from 10% to 100%), and the lack of solutions for accurate fruit estimations, all the players along the fruit chain are losing revenues.