112 Forage mass estimation of Cereal Rye in grazing systems using plant coverage analysis via smartphone images

Abstract

The current process of estimating forage mass requires significant time and effort. This project aims to improve pasture management by providing a tool that uses smartphone-taken images to overcome the difficulty of assessing the amount of forage available for grazing. Cereal rye forage mass data was collected by clipping forage to ground level within a 0.5 m2 area throughout two grazing seasons. Imagery data at 1 m above and parallel the ground was taken prior to clipping. These images were analyzed to calculate plant green canopy cover percentages using the Mahalanobis distance classification method. From May 2 to 23, 2023, forage mass in 124 areas was collected with a range of 87.4 to 4,806.21 DM kg/ha and a green canopy cover range of 2 to 80%. From April 1 to May 24, 2024, 276 biomass areas were collected, with a range of forage mass of 98.6 to 4,341.05 DM kg/ha, and green canopy cover range of 3 to 65%. A regression analysis between the amount of green cover and biomass (kg/ha) was performed in SAS using the PROC GLM procedure. The two grazing seasons dataset was then split into training (80% of the data) and validation (20% of the data) datasets. The estimation performance (predicted mass versus measured mass) was performed by applying the quadratic equation obtained from the training dataset to the plant coverage values in the validation dataset, then estimation performance was assessed using the PROC REG procedure in SAS. The quadratic regression of measured forage mass and green cover had an adjusted R² of 0.76 (P< 0.0001). When this equation was applied to the validation data set, the R2 between the predicted vs. measured biomass was 0.77 (P< 0.0001), with a RMSE of 423 DM kg/ha. To conclude, the plant cover image analysis tool proved to be promising for biomass estimation of cereal rye. This method can save time and allow producers to determine appropriate carrying capacity.