A Model for Early Detection of Paddy Leaf Disease using Optimized Fuzzy Inference System

Now-a-days disease is one of the most significant issues in the agricultural field especially in the paddy leaf because it gradually minimizes the productivity with a degradation in the health condition of the rice. The issue present in the agricultural field is reduced by different image processing and soft computing approaches but in certain situation, the elimination of the disease is still remains as a bottleneck. Hence, in this paper novel automatic paddy leaf disease detection using optimized fuzzy interference system (OFIS) has been proposed. Initially, the captured paddy images are transformed into Red, Green and Blue band and noise present in the green band is removed with the help of median filter. Afterwards, the texture and colour features are extracted from the pre-processed green band. Then, the extracted features are given to the OFIS system to classify the image as normal or diseased. FIS is a rule based algorithm and it used linguistic variables for classification process. To enhance the fuzzy system, the parameter of fuzzy system is optimally selected with the help of variable step size firefly algorithm (VSSFA). The outcome of the proposed system is analyzed in terms of Accuracy, sensitivity, and specificity. © 2019 IEEE.