IBM announced the Food Waste Developer Challenge last April, asking hackathon competitors to identify and build a novel solution to the problem of food waste using IBM’s own open-source technologies. This week, the winner was announced: FreShip.
FreShip aims to cut food waste by rerouting food from one source to another, based on how fresh the food is, as determined by a computer vision algorithm. It’s based on the idea that different buyers will have different requirements or expectations when it comes to how fresh they want particular foods to be. A supermarket might want to stock fruits before they even ripen whereas a baker might want them over-ripe for a better flavor in their baked goods.
FreShip’s machine learning solution is designed to visually analyze the relative freshness of a foodstuff or set of ingredients and then reroute those foods through the supply chain so none of them go to waste. The visual analysis is handled by IoT sensors of various kinds being integrated into shipping containers.
Using the same fruit analogy, if currently a supermarket receives fruit too ripe to sell, the loss might be written off and the food goes to waste. With FreShip’s algorithm, those same fruits would simply be resold and delivered to a new vendor who would make better use of them.
It’s a fairly simple yet elegant solution, and it could have a major impact on supply chains and food waste.
Currently, according to the US Department of Agriculture (USDA) between 30-40% of fresh food is simply wasted at one point or another of the supply chain. This is a massive problem as it’s not just food that’s wasted, it’s the labor, energy, water and time that went into its production that’s written off, too. Making smarter, real-time decisions about where food should be rerouted to, in order to minimize wastage, could put a sizable dent into that 30% wastage figure.
Given that companies in general are moving towards greener, cleaner and more efficient technologies, even outside of the competition FreShip might be onto a real winner.