An Identification of Guava Fruit Disease Using ML

Fruits contain a lot of fiber. Fruits contain biologically active substances that improve our health. This study focuses on identifying and classifying guava fruit diseases. Guava disease has become a significant issue for guava production, threatening farmers' socioeconomic growth. As a result of this phenomenon, an automated computer vision-based guava disease diagnosis system was developed that can detect malicious guava and direct it to early cure approaches, reducing relative economic loss. In light of this, we present in this study a guava illness detection and categorization recommendation approach based on convolutional neural networks (CNN). This paper employs transfer learning to classify images using our customized convolutional neural network (CNN) model architecture. The photographs in the data set were obtained from the Mendeley Data website. The data set includes 306 images of healthy and unhealthy guava fruits and leaves. The four disease classes targeted by the data sets are dot, Canker, mummification, and Rust. We obtained an average accuracy of 74.72% for training and 34.74% for testing from the data set used.