An Image Segmentation Technique -OEM for Plant Leaf Disease
Shanmuga Rajathi D1, D. Maheswari2
1Shanmuga Rajathi D *, Research Scholar, School of Computer Studies –PG, Rathnavel Subramainam College of Arts & Science, Sulur, Coimbatore.
2Dr. D. Maheswari, Head & Research Coordinator, School of Computer Studies –PG, Rathnavel Subramainam College of Arts & Science, Sulur, Coimbatore.
Manuscript received on January 09, 2020. | Revised Manuscript received on January 22, 2020. | Manuscript published on January 30, 2020. | PP: 2842-2846 | Volume-8 Issue-5, January 2020. | Retrieval Number: E5759018520/2020©BEIESP | DOI: 10.35940/ijrte.E5759.018520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Image segmentation plays a vital role in identifying plant leaf diseases. Hence it is considered as categorizing of a test image as set of non-continuous regions which are varied according to the features and its characteristics of the image along its properties in terms of homogeneous and computation on the grey level, texture and color component to provide easy image analysis. Familiar existing techniques for leaf disease segmentation use watershed method, thresholding and region based method. One applying these techniques, particular lesion represents a varied shape, texture and Color properties which makes the complex in the segmentation. In addition, these methods face several challenges such as inhomogeneous object detection and fragmentation. To combat those challenges, a segmentation model named as Object Evolution Mapping (OEM) has been proposed in this paper. It is developed for discretized representation of the inhomogeneous object based on the weight probability with specified limits. The disease affected area is considered as object, as affected region may appear in varied shape and texture, the proposed model strongly correlate those changes through error correction process. Furthermore abstraction building has been carried out by the objective function on the matrix for the determine the correlation of the pixel based on the shape and texture interpretation on the image. It extracts the inhomogeneous objects accurately by traversing the horizontally and vertically. Finally changes between the object is computed accurately on the each positions as pipeline procedure. Experimental results show that proposed OEM model provides the good result in terms execution time and accuracy on comparing it with existing approaches.
Keywords: Plan Leaf Disease, Image Segmentation, Error Correction, Shape and Texture Analysis.
Scope of the Article: Image Analysis and Processing.