Smart Irrigation Control With Disease Detection Using Machine Learning and Arduino Controller

Abstract

Our paper proposes an approach for the classification of leaf disease, based on the characterization of texture, shape, and color properties. An original plant leaf is preprocessed initially using the Gaussian filter to minimize the noise. For enhancing the contrast and quality of the image, the histogram equalization applied. The preprocessed image is segmented by K-means clustering; only affected region is picked and their features are extracted. The GLCM and the LBP systems are introduced for the extraction of features. It has the issues of lower accuracy and recognition rate. The proposed feature extraction techniques overcome the difficulties faced by the existing method. The feature comprises of texture features, shape features, and color features. Then, classification algorithm is applied over the segmented image in order to predict and classify the disease. Atlast, it compared with training images in relevant to show the performance assessment of the proposed approach. Based on disease classification, pesticides will be providing to plant leaf; Sensors are also connected to identify the water level, humidity those depends on the threshold value the motors will be on.