Development and Evaluation of a Yield Monitor for Round Hay Balers

Corn, grain and cotton yield monitoring technologies have been widely implemented since their development and throughout the past twenty years. Yield monitoring has been indicated to be the second most applied precision agriculture technology, behind auto-steer. However, commercially available technologies are for yield monitoring crops other than corn, cotton, and grain have not been widely available, if available at all. Yield monitoring for hay and forages has not been broadly implemented but it can be a beneficial management and inventory tool for farmers in the cattle and hay industries. This research focused on development and testing of a yield monitor installed on a John Deere model 458 round hay baler, but adaptable to any type of hay baler. The yield monitor was built using on-the-go remote sensing technology to measure the height of the windrow as it was being collected by the baler. Bales were individually weighed using a platform placed on truck scales, and samples from each bale were oven dried to calculate dry weights. For year 1, sensor data across 77 bales of a mixed hybrid bermudagrass (Tifton 85 and Coastal), 23 bales of Tifton 85 bermudagrass, and 9 bales of alfalfa were used to develop regression models predicting bale counts, baled wet weights, and baled dry weights. Data analysis indicates that yield prediction error, when applying the proper regression model, ranges between 3% and 15%. For year 2, 21 bales of mixed hybrid Bermuda (Tifton 85 and Coastal) and 20 bales of Tifton 85 bermudagrass were analyzed to formulate prediction errors ranging between 5% and 9.5%. The technology developed under this project provides the capability of generating hay yield maps with an acceptable level of accuracy for guidance in making variable rate prescriptions and zone management applications across a variety of hay crops.