Smart Agriculture Framework for Automated Detection of Leaf Blast Disease in Paddy Crop Using Colour Slicing and GLCM Features based Random Forest Approach

In the era of technological advances, the food needs of the huge population of the world could be tackled by adopting modern technologies like internet of things, artificial intelligence and cloud computing. These technologies can help in increasing the yield of food grains through precision agriculture. Paddy is one of the major food grains in the world, whose yield is mainly affected by leaf blast disease. This disease should be identified at an early stage in an automated way to mitigate its adverse impact on the yield of paddy crops through appropriate actions. In this concern, the present work proposes an automated framework for the accurate detection of blast disease affected leaves among healthy leaves of paddy plants through image processing and machine learning. This work makes use of colour slicing and grey level co-occurrence matrices (GLCM) procedures for processing and texture-related feature extraction of leaf images of paddy crop. These features are fed to machine learning classifiers viz., random forest, decision tree, kNN, XGBoost, AdaBoost and Histogram based gradient boosting algorithm. The thorough analysis of these classifiers through several performance measures reveals that random forest provides maximum accuracy of 99.10%, sensitivity of 99.05% and specificity of 99.05% for identification of leaf blast disease in paddy crops. Thus, colour slicing and GLCM features-based random forest classification approach is an appropriate choice for futuristic internet of things enabled automated framework to precisely detect blast disease affected leaves among healthy leaves of paddy crop.