Fruit Disease Detection Using Color, Texture Analysis and ANN

Abstract

Efficient growth and increased fruit yield are required and significant now that prohibitive demand exists for agricultural industry. To this end, farmers need manual fruit monitoring from harvest to development. However, manual supervision will not always yield adequate results and specialist guidance is always needed. It therefore involves the proposal of a smart agriculture technique that contributes with less human effort to better yields and development. We are introducing a method for the diagnosis and classification of foreign diseases in fruit.The traditional scheme uses thousands of vocabulary leading to language boundaries. Whereas the framework we developed uses image processing techniques as a simple way to express the image. The proposed work is carried out using the OpenCV library. In the image segmentation, the K-means cluster approach is applied; the images are catalogued on four vectors of color, morphology, form, and hole configuration based on their respective disease categories. The framework uses two image libraries, one for query image execution and the other for the training of previously stored pictures.The definition for pattern matching and disease recognition is used by Artificial Neural Network (ANN).