Multispectral images for estimating morphophysiological and nutritional parameters in cabbage seedlings

Abstract

Remote sensing data have been used to monitor numerous agricultural crops, but few studies have used low-cost multispectral sensors to assess biometric, nutritional and physiological parameters in a cabbage crop, especially in seedlings. As such, this study aimed to estimate the biometric variables of cabbage using parametric and non-parametric models based on the response of multispectral images taken by a multispectral camera. The experiment was conducted in a greenhouse in the municipality of Uberaba, Minas Gerais state (MG), Brazil. Twenty days after sowing (DAS), multispectral images of the plants were captured using a MAPIR Survey 3 camera. To compose the estimation models, the multispectral vegetation indices were calculated based on the original calibrated bands. Multispectral cameras can be used to estimate biometric, physiological, and nutritional variables in cabbage crops. Among the Mapir camera bands, B660 exhibited the greatest variability, showing that the red range was the most sensitive to the different treatments. Except for leaf area, all the parameters measured could be estimated by linear models, with deterministic coefficients of up to 72%. Although neural network models provided more accurate estimates, the maximum quantum yield of photosystem II (Fv/Fm) was most accurately estimated (1.2%) by the linear model generated by algorithm M5.

Keywords

• Brassica oleracea var. capitata
• Morphobiometrical variables
• Low cost multispectral camera
• Prediction models