Adaptive neuro-fuzzy inference system for identification of rice plant diseases

The causes of decreased rice productivity include rice plant diseases. Therefore, identifying diseases that attack rice plants must be done to treat these diseases. So that rice plants are free from disease in the future. Currently, identification of rice plant diseases can only be performed by human agents who possess expert knowledge. In this study, we carried out a disease identification method in rice plants based on artificial intelligence and image processing without human intervention. This new method begins with data collection, image pre-processing, image segmentation using k-means, and then extracting color and texture features. The final process is the classification of rice plant diseases using the Adaptive Neuro-Fuzzy Inference System (ANFIS) method. Experiments were carried out on multiple leaf images, splitting them into three classes: blast disease, brown spot disease, and healthy. The highest accuracy value obtained is 86.67% using 11 features, which suggest that our new method is reliability and of a high accuracy in the identification of rice plant diseases.