A Novel Mining Approach for Automatic Disease Detection in Sugarcane Plant Using Thresholding Method

Abstract

Disease infection of agricultural products affects the agriculturists and human health and also degrades the quality and quantity of products. The tradition approach for detecting a disease is time consuming and very costly. The plant diseases are detected by automatic detection techniques that reduce a large work of continuous monitoring and observation in big farms by farmers or experts. The proposed algorithms detect the variety of diseases infected in sugarcane plants. The images are captured by digital camera. The noises in digital image are removed by low pass filter. This paper presents image segmentation using thresholding method which is used for automatic disease detection of sugarcane plants. SIFT method is applied for detecting and describing the local features of the plant species. The features such as colors, size shape and texture of surface is extracted by using GLCM feature extractor. The abnormal images are classified by using SVM classifier. In sugarcane plant, the diseases are detected automatically and yields 99% accuracy rate than existing techniques.