From Farm to Market More Efficiently
When it comes to agriculture, weather isn’t all that’s dicey. Farmers need accurate counts of the produce in their fields to determine pricing, negotiate contracts with buyers, and maintain the overall financial health and success of their farms. But precise crop estimation techniques are expensive to execute because they entail high labor costs, while more efficient estimating techniques are often inaccurate.
Enter Penn Engineering student Wei-Yin Ko, GEN’18, who’s working on a master’s degree in robotics. In the course of his research, he examined precision agriculture, a term that encompasses everything from GPS soil sampling to videotaping drones that fly over crop fields. He also studied a machine-learning method of estimating crop data by relying on images.
In his work, Wei-Yin witnessed firsthand how research is often bottlenecked by a lack of data or by labeling methodology that’s been developed for onetime applications by researchers themselves. The result is a lack of efficiency: Research time is spent either acquiring more data or converting data formats. This inefficiency extends to corporate ventures, since most machine-learning algorithms are developed in-house.
On a school-break visit with friend George Bumgardner to North Carolina, Wei-Yin met local melon farmers and discussed his research with them. That’s when he discovered the need for more accurate yield estimates and realized that he could utilize computer learning to automate the process, improving precision and efficiency.
The result is AgGrow, a service that scans images provided by farmers, taken either with their cellphones or by drones, and accurately assesses crop counts (and, eventually, will provide additional predictive information for free). AgGrow is already saving time and money for farms of all sizes and improving the efficiency of the entire agriculture industry. The product garnered the team —Wei-Yin, George, and their partner Ben Rhodes — the Best Use of Technology Award in Penn Wharton Entrepreneurship’s Startup Showcase as well as semifinalist standing in the Startup Challenge in 2017.