An Empirical Investigation of Olive Leave Spot Disease
Using Auto-Cropping Segmentation and Fuzzy C-Means Classification
Mokhled S. Al-Tarawneh
Department of Computer Engineering, Faculty of Engineering,
Mutah University, P.O. Box: (7), Mutah 61710, Jordan
Submitted: Jun 8, 2013; Accepted: Jul 15, 2013; Published: Jul 20, 2013
Abstract: The objective of this research was to investigate an image analysis and classification techniques for detection and severity rating of olive leaf spot disease. Samples of olive leaves were collected from field and imaged under uncontrolled illumination. Images resolution were resized to 256×256 pixels and transformed from RGB to L*a*b* color space. The transformed images were then cropped polygonal to segment the region of interest and classified using fuzzy c-mean clustering for statistical usage to determine the defect and severity areas of plant leaves. Imaged enhancement was performed using median filtering. The severity percentage was calculated based on classification of detected diseased and total leaf areas. Comparative assessment of FCM and KCM was conducted with reference to mean opinion scoring of image data. The results showed a good agreement between FCM and manual scoring and by image analysis at an 86% accuracy rate comparing to KMC with 66%.
Key words: Olive leaf spot disease Cropping Color transformation Segmentation and Image classification