期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2017
卷号：6
期号：8
页码：15696
DOI：10.15680/IJIRSET.2017.0608024
出版社：S&S Publications
摘要：Agricultural yield is mainly affected by differences in climatic conditions and seasonal diseases. There isa strong need of modern agricultural systems that identify disease in the crops. The goal of this study to design, developand evaluate GA-SVM based methodology for automatic classification of pomegranate related diseases. Thisinnovative method would be a boon to many and would have lot of advantages over the traditional method of grading.Presently, plant pathologists follow a tedious technique that mainly relies on naked eye prediction. Manual procedure isnot only time consuming but also does not give accurate results. Studies show that relying on pure naked-eyeobservation of experts to detect and classify pomegranate related diseases can be prohibitively expensive, especially indeveloping countries. This methodology is fast and accurate in detecting diseases like bacterial blight, cercosporal leafspot and wilt of pomegranate. Present experimental results indicate that the proposed approach can significantlysupport an accurate and automatic detection and recognition of leaf diseases. Current paper proposes GA-SVM basedmethodology to deal with one of the main issues of plant pathology. The results are proved to be accurate andsatisfactory in contrast to manual procedure and hopefully take a strong leap forward in establishing itself in the marketas one of the most efficient and effective process. In conclusion, the proposed detection models based GA-SVM arevery efficient in detecting and classifying bacterial blight, cercosporal leaf spot. The developed Genetic Algorithmbased SVM classifier that based on statistical classification perform well in all sample types of leaf diseases and cansuccessfully detect and classify the examined diseases with accuracy 94%.
关键词：Leaf diseases; cercospral leaf spot; bacterial blight and texture attributes.
