Early Phase Detection of Bacterial Blight in Pomegranate Using GAN Versus Ensemble Learning

Bacterial blight is a severe disease that impacts pomegranate trees, leading to significant yield losses. This study proposes an innovative approach to detect bacterial blight in the early stages using image processing techniques. The method involves capturing images of pomegranate leaves using a digital camera and analyzing them using various image processing algorithms, including segmentation, feature extraction, and classification. This non-invasive and cost-effective method accurately detects bacterial blight, allowing farmers to take preventive measures and minimize yield losses. The suggested method can become a beneficial resource for farmers and researchers involved in plant pathology. Using ensemble learning, pomegranate bacterial blight in its early stages was discovered. The study findings showed that the proposed approach was more accurate and had a higher detection rate than traditional methods. Without the use of paired training data, the suggested method of Cycle-GAN is a sort of Generative Adversarial Network (GAN) that can learn to translate images from one domain to another. Cycle-GAN has demonstrated its efficacy in various image-to-image translation problems by utilizing the potential of unsupervised learning.