A fruit that’s ready to harvest tends to weigh down the branch, making it easier to spot and easier to reach. But what happens when a fruit grows so low to the ground, in such densely packed blocks, that you can’t easily walk among (or under) the plants? How do growers determine which fruits are ready for harvest, and which ones aren’t? Case in point: the natural flowering of pineapple was the basis of the industry up until the 1960s. Now, pineapple fields are forced in blocks to flower, with a chemical that releases ethylene and induces flowering, making the fruit available year-round.
However, if the days are short and the nights cool, natural flowering will occur potentially before a scheduled ethylene treatment, leading to two flowering peaks and hence, two fruit maturation peaks and more than one harvest. Since pineapple is hand-harvested, a grower’s ability to harvest all of the fruit of a field in a single pass is critical to reduce field losses, costs, and waste, and to maximize efficiency.
In a new study funded by the USDA’s Small Business Innovation Research program, researchers in CTAHR, IntelinAir, Inc., and Columbia University are investigating whether remote sensing and computer vision can help growers carry out regular inspection of the field and automated counting of flower intensity.
“Our work used deep learning-based density-estimation approaches to count the number of flowering pineapple plants in a field block,” says Robert Paull of the Dept. of Tropical Plant and Soil Sciences.
“This enables growers to optimize their planning and management practices for getting optimum fruit harvest.”
Drones, he explains, are being used worldwide to monitor crop growth, disease, and weeds, and to apply fertilizer and crop protection products. The tool allows growers to be more flexible, efficient, and highly targeted, with lower costs and input application. They’re also able to service hard-to-reach areas and where weather prevents access by heavy equipment.
“New technology and management strategies are critical for the economic success of farming in Hawaii,” Robert adds. “Drones are used widely, though less effort has been devoted to tropical agriculture systems. In the tropics, drones offer the ability to enhance precision agriculture, improve crop management, and reduce environmental impacts and costs.”
Read the full study, Large-Scale Counting and Localization of Pineapple Inflorescence Through Deep Density-Estimation, which appears in a recent edition of Frontiers in Plant Science.