Predicting Quality And Quantity In Apples

What is the possible connection between the tragedy of the BP Deepwater Horizon spill and tree fruit crops? I suppose there might be quite a number, but the one that strikes me is how difficult it really is to measure stuff that is important.

One would think that a company like BP would have enough engineering expertise and specialized equipment to accurately estimate the oil and gas flow from the ruptured line. We know exactly where it is and can even watch it live on the Internet.

We might say the same thing of tree fruit crops. We know exactly where the fruit is in an orchard from flowering though harvest.

In the BP spill, there is considerable public incentive to accurately estimate the quantity and flow of the oil and gas. For tree fruit producers, warehouses, and marketing organizations, it is not that different. They want to accurately estimate the quantity of fruit on the trees and their growth rate.

Accurately predicting the quality and quantity of fruit is critical from production, storage, and shipping activities all the way to point of sale. But, is it really possible for us to get an estimate in July for the 10 billion apples Washington typically packs for the fresh market on an orchard-by-orchard basis?

Scouting And Counting

Even if we could recruit enough people to scout and count fruit all summer long, how could we pay them? Since we can’t, and since this need for more accurate crop load estimation is so critical, a number of industries are working with research groups to develop quick and efficient sampling schemes, principally focused on automated approaches based on various types of sensing technologies.

One of those approaches is being field-tested in the Western U.S. this season, on oranges in California and apples in Washington. This project, led by Vision Robotics, features a mechanical scout. The prototype is towed by an ATV and uses cameras operating in the human range of vision, a specialized lighting apparatus, portable computers, and sophisticated custom software.

After three years of research, we believe the scout is now capable of rapidly scanning portions of an orchard, processing the massive dataset, and reporting with precision the number of apples it has seen. Eventually, we expect it will not only count the number of fruit but accurately estimate fruit size.
If we can run the scout through the orchard during the season and ascertain growth rates, we will then be able to predict crop size and the characteristics of individual apples at harvest.

Further Research

At present, the scout requires a human operator. However, as part of a project funded through the Specialty Crops Research Initiative (SCRI), researchers are working to transform the scout into a fully autonomous unit. In the same project, researchers from Penn State and Washington State Universities are collaborating to adapt these technologies to further automate orchard operations from pruning to thinning to spraying to harvest.

Another SCRI-funded project, Integrated Automation for Sustainable Specialty Corps, involving engineers from the National Robotics Engineering Center and the University of Florida is taking a different, complementary approach with similar objectives on citrus. Yet one more SCRI project, focused on sweet cherries and led by scientists at Washington State University, includes some similar activities.

These projects represent a new wave of research bringing engineering solutions to the problems faced by specialty crop producers. As the mantra goes, “If we can measure it, we can manage it.” No doubt the millions expended on measuring the BP spill flow rate is important, as well. I can’t help but wish we were in a position to use them on a project improving agricultural sustainability rather than mitigating an environmental disaster.