Geometry based hybrid method for determining lesion area

Abstract

In this paper, a hybrid method to measure lesion area on the green leaf surface is presented. The method involved combining two algorithms, Artificial Bee Colony and Otsu algorithm and applying geometry to obtain the area. Three leaf images were used that contains leaf lesions of different sizes: small, medium and big. The method was conducted in three phases, data preparation, algorithm construction and analysis. Comparison was made with pixel counting, and real measurement. The results show that the proposed hybrid method achieved results more accurate (average accuracy 92.03%) and faster (average time processing 0.015 milliseconds) than the pixel counting method.