A Tree Image Segmentation Method Based on 2-D OTSU in HSI Color Space

This paper concerns the problem of tree image segmentation. Image segmentation is the basis of the precision toward-target pesticide application to trees and the real-time and accuracy requirements are very high. In this work, we transformed the image color space from RGB to HIS and normalized the image hue centered on tree hue. Then the color image was segmented by using of an improved 2-D OTSU method. Results show our method has good accuracy and real-time performance and can be used in the precision toward-target pesticide application to trees.