RGB-based indices for estimating cover crop biomass, nitrogen content, and carbon:nitrogen ratio

Authors: ROSEN, LUCAS - University Of Minnesota; Ewing, Patrick; RUNCK, BRYAN - University Of Minnesota

Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/19/2024
Publication Date: 9/21/2024
Citation: Rosen, L.B., Ewing, P.M., Runck, B. 2024. RGB-based indices for estimating cover crop biomass, nitrogen content, and carbon:nitrogen ratio. Agronomy Journal. 116(6):3070-3080. https://doi.org/10.1002/agj2.21657.
DOI: https://doi.org/10.1002/agj2.21657

Interpretive Summary: Nitrogen-fixing cover crops provide many agronomic benefits to grain production systems, including weed suppression, forage production, and nitrogen credits. Managing them efficiently within cropping systems requires estimating these benefits using such metrics as canopy coverage, biomass, and nutrient content, respectively. We tested image processing methods that used a consumer-grade RGB camera to estimates of these important metrics. We studied medium red clover planted under an oat cash crop or after oat harvest and took images in highly uncontrolled field conditions. The top method, the excess green minus red index combined with a plant/not plant index threshold of zero, was able to identify plants in images with 86% accuracy at the pixel level. Those pixels identified as plant could further be used to estimate biomass (RMSE error = 220 kg/ha) and nitrogen content (RMSE = 0.34 percentage points). Pending testing across different sensors, sites, and crop species, this method contributes to a growing, open ecosystem of decision support methods for highly distributed and farmer-led agricultural research and paves the path toward more efficient and sustainable grain production.

Technical Abstract: Plant cover and biochemical composition are essential for the research and management of cover crops. Destructive sampling or estimates with aerial imagery require substantial labor, time, expertise, or instrumentation cost. Using low-cost consumer and mobile phone cameras to estimate plant cover and biochemical composition could broaden the use of high-throughput technologies in research and crop management. Here, we estimated canopy development, tissue nitrogen, and biomass of medium red clover (Trifolium repens L.), a perennial forage legume and common cover crop, using common non-expert settings in field conditions. Pixels were classified as plant using combinations of four red-green-blue (RGB) indices with both unsupervised machine learning and preset thresholds. The best combination, the excess green minus red (ExGR) index with a preset threshold of zero, correctly identified pixels as plant or background 86.25% of the time. Critically for research and management, this combination also provided accurate estimates of crop growth and quality: Canopy coverage correlated with red clover biomass (r2=0.569, RMSE=219.29 kg/ha), and normalized green values of leaf tissue pixels were highly correlated with clover nitrogen content (r2=0.573; RMSE=0.34 percentage points) and carbon:nitrogen ratio (r2=0.596, RMSE= 1.29). Data collection was trivial to implement and robust to imaging conditions. Pending testing across different sensors, sites, and crop species, this method contributes to a growing, open ecosystem of decision support methods for agricultural research and management.