Work

Development of an in-season nitrogen application dose estimation algorithm for cotton using multispectral imaging-based nitrogen adequacy index

Year: 2026

Type: article

Abstract: Managing within-field variability in cotton fields for precision nitrogen (N) management is difficult. The development of multispectral sensors and image data analytics offers a solution to addressing the issue. During the Kharif (monsoon) season of 2021 and 2022, field experiments were conducted at the ICAR-Central Institute for Cotton Research, Regional Station, Coimbatore, Tamil Nadu, involving seven N levels (25%, 50%, 75%, 100%, 125%, 150%, and 200% of the prescribed N-dose (80 kg ha −1 )), along with a control (N0). Unmanned aerial systems-based multispectral crop imaging data were collected to develop an algorithm for calculating the N-dose for in-season variable-rate application. The multispectral images were processed for different crop growth stages, and treatment-wise mean normalized difference vegetation index (NDVI) and normalized difference red edge (NDRE) values were determined. The relationship between the mean NDRE index and estimated leaf N content exhibited statistically significant regression at 70 days and 95 days after emergence (DAE) during Kharif 2021 and 75 DAE and 90 DAE during Kharif 2022, respectively. Spatial nitrogen adequacy index (NAI) maps were created using the 95th percentile NDRE values (0.37 for Kharif 2021 and 0.27 for Kharif 2022). The N rate that maximized cotton yield was used to parameterize the NAI-based N-sufficiency response curve. The quadratic model (N rate (kg ha −1 ) = −191.25 (NAI) 2 −22.32(NAI) + 138.5) produced the best fits, with a coefficient of determination ( R 2 ) value of 0.84. The developed algorithm can be used in preparing N zone maps for variable-rate N application and has the potential to optimize cotton’s in-season N requirement.

Source: Frontiers in Remote Sensing

DOI: 10.3389/frsen.2026.1779561