Visual Features based Paddy Leaf Disease Recognition, its Severity Detection and Remedy Prediction using K-means Clustering and AdaBoost
Agriculture is a vital part of an economy. Paddy is one of the main food crops which play a major role in agricultural field. The gross national income of a country depends on paddy cultivation. But the production of paddy is damaged due to different types of paddy leaf diseases. Generally, farmers and agricultural experts identify diseases manually which is very ineffective and time consuming. So effective recognition, severity detection, and proper management of paddy leaf diseases are necessary. This paper presents visual feature based recognition system for three common paddy leaf diseases (Brown Spot, Leaf Blast, and Bacterial Leaf Blight) using K-means clustering and AdaBoost classifier. To separate the affected cluster from diseased leaf, K-means clustering is used, and the affected ratio is calculated to detect the severity of diseases. The classification of diseases is performed using color and texture feature analysis. Mean and standard deviation are used as color feature. For texture feature extraction, correlation and wavelet packet entropy are used. AdaBoost classifier is applied to recognize the type of diseases. After recognizing the diseases, necessary remedy is suggested so that farmers can take necessary measures for management of paddy leaf. The proposed system shows comparatively robust result.
Keywords:AdaBoost, affected ratio, color moments, correlation, K-means clustering, remedy, wavelet packet entropy
- There are currently no refbacks.