Automatic image segmentation of greenness in crop fields

Abstract

This paper describes a new automatic image segmentation strategy for segmenting green plants. The final goal is its application in Precision Agriculture. The goal is to identify several classes of greenness coming from the plants. We exploit the performance of several existing approaches so that conveniently combined allow us to design the automatic approach based on non automatic methods. First we apply a well known index-based strategy that accentuates the green spectral band from the remainder, giving a gray image. From the resulting image we apply the well-known thresholding Otsu's method obtaining a binary image, where the green part appears separated from the soil. Taking as input the green pixels we apply an unsupervised method and they are partitioned in a fixed number of classes. The performance of the method is tested against a set of available images and acquired in several crop fields of cereal and maize.