New method for cotton fractional vegetation cover extraction based on UAV RGB images

As the key principle of precision farming, the distribution of fractional vegetation cover is the basis of crop management within the field serves. To estimate crop FVC rapidly at the farm scale, high temporal-spatial resolution imagery obtained by unmanned aerial vehicle (UAV) was adopted. To verify the application potential of consumer-grade UAV RGB imagery in estimated FVC, blue-green characteristic vegetation index (TBVI) and red-green vegetation index (TRVI) were proposed in this study according to the differences of the gray value among cotton vegetation, soil and shadow in the field. First, two new constructed indices and several published indices were used to extract visible light images and generate greyscale images for each of the visible light vegetation indices. Then, the thresholds of cotton vegetation and non-vegetation pixels were established based on the vegetation index threshold method which combines support vector machine classification and vegetation index. Finally, the accuracy difference in vegetation information extraction between the newly constructed and several published indices was compared. The results show that the accuracy of the information extracted by TRVI is higher than that of subdivision index of other visible light (FVC extraction precision in the first bud stage of cotton: R-2=0.832, RMSE=2.307, nRMSE=4.405%; FVC extraction precision in the bud stage of cotton: R-2=0.981, RMSE=1.393, nRMSE=1.984%; FVC extraction precision in the flowering stage of cotton: R-2=0.893, RMSE=2.101, nRMSE=2.422%; FVC extraction precision in the boll stage of cotton: RV=0.958, RMSE=1.850, nRMSE=2.050%).