International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 07 | July 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 4460 Implementation of Plant Leaf Diseases Detection and Classification using Image Processing Techniques Chandan Kumar Singh 1 , Dr. Sandeep B. Patil 2 , Dr. Om Prakash Sahu 3 1* (M.Tech, Student), 2* Associate Professor, 3* Ass.Professor Department of Electronics & Telecommunication Engineering, Shri Shankaracharya Group of Institutions (SSTC) Junwani, Bhilai, C.G, India. Vellore Institute of Technology, Chennai -----------------------------------------------------------------------***------------------------------------------------------------------------ ABSTRACT: In the proposed system we can detect the plant leaf disease and identify the disease. Captured Images and data base are maintained with the sample size of leaf. By using SVM classifier method, recognizes and detect the percentage of affected region, types of disease with the accuracy of 98%. Via Alternaria Alternata, Anthracnose, Bacterial Blight and Cercospora Leaf spot detection SVM classifier technique is used to identify the plant leaf disease . Keywords: Leaf Image; Image enhancement; Image contrast; Support Vector Machine; 1. INTRODUCTION Automatic detection of plant leaf disease and classification of leaf disease is the most important research topic as it shows an advantage in monitoring of large area of field crops and routine detect and classify the plant leaf disease. Leaf image can be capture by using the mobile camera of 25 Mega pixels or above and leaf images can be detected and classify by using the SVM classifier. With this method farmers can easily identified the disease and according preventative measures can be taken to cure the disease and take maximum crop production. In the existing system, we are inspection the factors of the plan manually. It needs the physical inspection of plants through the naked eye and time consuming to inspected the whole field. Continuous observation of crops and maintenance is incredibly troublesome. This could decrease in crop yield production because of poor observance. Drawbacks of manual monitoring Time consumption, Manpower requirement, Detection of disease through the naked eye 1.1 Proposed system: In this system images are saved in data base and by using the SVM classifier technique we can identify the percentage of affected part of the leaf and according proper method / pesticide/organic manure is used to cure the disease. This method gives the accuracy of 98%. Fig.1 Proposed System block diagram [1] Author proposed the Convolution Neural network (CNN) for detection of cotton leaf disease with the accuracy of 86%.[2] GLCM method is applied feature extraction and KNN classifier is used for detection with the accuracy of 95%.[3] Support Vector machine (SVM) algorithm is sued with five kernel function i.e., Linear, quadratic, radial basis function, sigmoid and polynomial. This method gives us the grayscale conversion. Classification was done on the basis of statistical, color and texture features based on SVM but accuracy level is 90% only.[4]Author proposed the k-means clustering algorithm and Otsu’s classifier in which the infected area of leaf is segmented and analyzed. [5]Author used the methodology of Simple linear iterative clustering (SLIC) is widely applied to super pixel clustering due to its simplicity and practicality. This proposed method is appropriate for dealing with plant disease leaf image segmentation and has certain superiority in the field of plant disease detection.[6]Genetic algorithms are used to set the unlabeled points in N-dimension into K cluster and for feature extraction color co-occurrence method are used. The minimum distance criterion with K-means clustering gave an accuracy of 86.54% and with SVM the accuracy was 95.71%.[7] Author proposed the statistical and structural recognition method. The statistical recognition of patterns totally depends upon the pattern characteristics which are also statistical in nature. Structural recognition of characterizes depend on the interrelationship among the structure which contain features.[8] For detection of plant leaf are BPNN,K-means clustering and SGDN are used . For classification of plant leaf disease SVM technique is implemented.[9] Author used the Vision-based detection algorithm with TYPES OF DISEASES MAT LAB IMAGES