Automatic Citrus Fruit Disease Detection by Phenotyping Using Machine Learning

It is really a frustrating occurrence of how fruit diseases can cause a reduction in production and the economy in the agricultural field all over the world. A growing body of research proves that fruits are critical to promoting good health. In fact, fruits should be the foundation of a healthy diet. This paper presents a better and modern proposed solution for the detection of fruit diseases using their physical attributes. The proposed methods are composed of K-Means clustering technique, ANN and SVM. K-Means is used for the image Segmentation. It has a function of mapping images to their corresponding disease classes based on the phenotypic characteristics such as the texture, color, structure of holes on the fruit and physical make-up. ANN (Artificial Neural Network) is pragmatic in achieving enhanced results in relations to the accuracy of detection and classification have some advantages over the other algorithms. They utilize quite little pre-processing in regards to other image classification methods. It implies the filters were studied by the network that in traditional methods were hand-engineered. This autonomy from earlier information and human exertion in feature design is a major advantage. A Support Vector Machine SVM also is used with the ANN to increase the high rate of classification. The proposed solution can significantly engineer precise detection and automatic classification of fruit diseases.