Studies on Image Recognition Algorithm for Cucumber Plant (Part 1)

Recognition of Cotyledon and True Leaf on Young Seedling Image

Abstract

An image recognition algorithm of plant was developed to automate management operations for cucumber plant considering plant growth rules. In this paper, a method to recognize cotyledon and true leaves of young cucumber seedling on binary image is reported. As features to discriminate cotyledon and true leaf, a complexity of boundary on a polygonal image and moment around leaf axis were extracted. It was observed that cotyledon and true leaf were identified with the probability higher than 92%. The identification from the whole image of seedling was possible with the probability of 89.8%, by verifying the discrimination results of respective leaves using leaf number and angle between leaves.