Feasibility Study Of Evaluation Brix Of Sugarcane Using Multispectral Camera Mounted On Unmanned Aerial Vehicle

Abstract: The possibility application of multispectral imagery using Unmanned Aerail Vehicle (UAV) as platform to predict Brix value in sugarcane field was evaluated. Total 4 flight missions were conducted in November 2017 using rotate- wing UAV and a 5- band MicaSen- RedEdge camera sensor to fly in the afternoon time (around 2:00pm) over an experiment trail which consists of 4 varieties (KK3, K88-92, UT84-12, KK06-501). The sugarcane is new planted, in its 10-month-old with the range of brix 15-21 0 Brix, and grown under rainfed. After each flight, Brix measurement was done using refractometer with 2 random juice samples in each plot, total 32 samples in this study. Acquired Images were processed in Pix4D mapper software to create reflectance map of each band (Blue, Greem, Red, NIR, and RedEdge). Matlab was used to compute vegetation indices (RVI, NDVI, GNDVI, SRPIb, CIgreen, and CIrededge). Then the calibration equations are determined using simple regression between each vegetation indices and brix value. The result showed that GNDVI has the best correlation with Brix value, Standard Erro (SE) 0.45 0 Brix and R2
0.86, while the rest vegetation indices have higher SE 0.51-1.13 0 Brix. From preliminary result, it is possible to predict Brix value in the field from spectra of multispectral camera mounted on UAV. The accuracy of model can be improved by adding more data from various concerning environment conditions such as light and temperature, and more range of brix data because datas used in this study are still in early harvest season.

read the paper

Authors: Chanreaksa Chea

Associations: 1Faculty of Engineering, Khon Kaen University, Muang, Khon Kaen, 40002, Thailand