Leaf area measurement of selected vegetable seedlings using elliptical Hough transform

摘要

An image-processing algorithm using the elliptical Hough transform was developed to determine position, orientation, and leaf area of seedling leaves from top-view images. The algorithm significantly reduced the computational effort and memory requirement and was capable of identifying partly occluded leaves. Four varieties of vegetable seedlings, namely cabbage, Chinese mustard, edible amaranth (A. mangostanus Linn.), and edible amaranth (A. inamoenus Willd.), at various growth stages were used to test the efficacy of the measurement algorithm. For measurements of individual leaf area, the relative estimation errors were 8.2 ± 9.3%, 14.4 ± 11.7%, 27.5 ± 14.4%, and 16.0 ± 11.0%, respectively. For measurements of total leaf area, the relative estimation errors were 16.0 ± 6.4%, 17.0 ± 9.8%, 24.9 ± 8.1%, and 18.4 ± 9.4% in corresponding order. The sources of error were mainly due to tilting leaves and unsuccessful identification of small or severely occluded leaves of the seedling.