Abstract

An automatic fractional vegetation cover (FVC) estimation method based on image characteristics in an agricultural region was proposed in this study to remove the empiricism in determining the key parameters of empirical methods. The proposed method automatically determined the soil and vegetation lines in the two-dimensional space of the red and blue band reflectances, which involved an iterative soil and vegetation pixels selection procedure, and then estimated FVC of a pixel based on its distances from the soil and vegetation lines. The accuracy assessment using field survey data indicated that the performance of the proposed method (R2 = 0.69, RMSE = 0.072, Bias = 0.014) was comparable with several commonly used empirical methods. Therefore, it was indicated that the proposed method could effectively estimate FVC in the corn-dominated region.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call