COTTON CANOPY NDVI: REDUCING THE GROUND EXPOSURE EFFECT

摘要

Cotton producers aim for optimal nitrogen (N) fertilizer application rates that are correctly timed. Supplemental N is only used at the growth stage when and where it can be most effective. One management strategy monitors N levels using leaf spectral reflectance throughout the growing season. However, when the canopy is sparse or absent, a mobile implement-mounted spectral sensor is subject to influence from the exposed soil and stubble background. The results are misleading readings. This exploratory research project evaluated two vegetation index (VI) algorithms. The optimal soil-adjusted vegetation index (OSAVI) and normalized difference vegetation index (NDVI) values were selectively trimmed for improving a commercial NDVI sensing system's ability to discriminate between plant biomass and the ground surface. Varying plant populations, especially in early growth, can lead to erroneous real-time readings during mobile applications when passing over exposed ground surfaces. The goal of this study was to improve the real-time monitoring of cotton N status, both spatially and at the field scale. A large-scale field experiment statistically established a distribution of small contiguous plots in rows. A plot unit was a combination of one of three seeding rates, one of three cotton varieties, and one of four N application rates. Plant height, leaf N, and VI values were collected and analyzed. Both VI algorithms were found to reduce the varying plant population effect on NDVI. Plant population affected raw NDVI values throughout the season, which confirmed the effect of soil background exposure on NDVI values. Research findings suggest that after applying the two algorithms and comparing results with the non-filtered data analysis, both algorithms detected the differences due to variety, seeding rate, and N rate treatments. However, the N effect was detected earliest in the season by NDVI.