Evaluation of Red and Red-Edge NDVI From Ground-Based Active Optical Sensors for Corn Yield Prediction Over Two Years Under Different Soil Type Texture and Tillage Categories.

Ground-based active-optical (GBAO) crop sensors have been used successfully to predict crop yield when used during early growth stages. The objective of this study was to evaluate two GBAO sensors regarding their usefulness under North Dakota corn (Zea mays, L.) growing conditions. Thirty experimental sites were used in North Dakota to conduct N rate trials on corn during 2011 and 2012. All sites were designed as randomized complete blocks with four replications and six nitrogen (N) treatments; check (no added N), 45 kg N ha^-1, 90 kg N ha^-1, 135 kg N ha^-1, 179 kg N ha^-1 and 224 kg N ha^-1as ammonium nitrate applied by hand preplant within a week of planting.  Results were analyzed for all sites. The sites were segregated into high clay and medium textured soils as well as no-till and conventional tillage. The Greenseeker ® and Holland Crop Circle-470 ® were used to gather NDVI readings from each plot at both the V6 and V12 stages. An interior of each 6.1-m by 3.05-m plot was hand harvested and the yield was related with sensor measurements to build algorithm for in-season estimate of yield (INSEY). Four wavelengths 656, 670, 730, and 760nm, were used to measure the red NDVI and red edge NDVI. Five regression models, linear, quadratic, square root, logarithmic, and exponential were used to evaluate the wavelength for INSEY. The Red NDVI from the Greenseeker and the Holland Crop Circle sensor provided similar relationship to yield as the Red Edge NDVI from the Holland Crop Circle Sensor at the V6 stage. The Red edge NDVI was generally highly related to yield compared to red NDVI across the models, soil cropping system, and soil textures at the V12 stage.