Sugarcane Disease Classification Using Soft Computing Algorithms

Sugarcane plant diseases are caused by various insects, bacteria, viruses etc., To reduce the impact of the diseases and thereby increase sugarcane yield, early detection of diseases is required. Existing techniques for detection of plant diseases focus on a single type of disease which does not have a significant impact on increasing the sugarcane crop yield. We propose an efficient Sugarcane Disease Classification system (SDC) using soft computing techniques. The proposed system has four different modules; colour space conversion, histogram generation, feature extraction and classification. Sugarcane image is represented by two different colour models RGB and La*b*.Wavelet and contourlet based energy features from both RGB and La*b* colour models are extracted and finally classified using a neural network. The proposed SDC system is tested on 1000 sugarcane images affected with ten different diseases (100 images per category) by a 10-fold cross validation approach. Results show that the SDC system provides an average classification accuracy of 99.89%