Olive Spot Disease Detection and Classification using Analysis of Leaf Image Textures
Aditya Sinha, Rajveer Singh Shekhawat
Abstract
The olive tree is a highly beneficial fruit tree with the earliest known history of its plantation going back to 6000 years. The production of olive oil is facing a significant threat nowadays due to climate change and the spread of diseases. In this paper, olive disease analysis and classification using image processing techniques are done. Using image texture analysis of olive plant leaves, correlation in between the signatures of Neofabrea leaf spot disease and Peacock leaf spot diseases with some texture features is identified. Neofabrea leaf spot disease manifests itself as distinguishable spots, prominently circular but is quite similar to the peacock leaf spot disease. In the proposed methodology, the infected area is isolated using the histogram thresholding and k-means segmentation. Out of the two methods, k-mean segmentation offered higher accuracy. In the next step, texture analysis is applied using first to fourth order moments, which in turn helped us to identify the relationship of infection with one or more texture features. A strong correlation between infection area and texture features such as energy, homogeneity, entropy is identified, which also helped in classifying the two similar diseases.