Weed Center Detection in Lawn Field Using Morphological Image Processing

Abstract

A method on uniformity analysis made weed detection in lawn field possible. The method could detect weeds based on their surfaces which were more uniform in gray-scale values compared with lawn surface, regardless of their color. The result of uniformity analysis, however, was far from ready to be used as a target decision for weeding because it was a binary image with many noises and broken parts. Morphological image processing was applied to the result of uniformity analysis to obtain stronger and clearer binary images. In addition, the detected weed center location which was expected to correspond with the main root of the weed, was determined for minimum chemical weeding. The techniques on how to do that was also described in this paper.