Image Super Resolution Using Extensive Residual Network (ERN) for Orange Fruit Disease Detection

Oranges hold global significance in trade and consumption. Diseases like citrus greening threaten orange crops. Timely disease identification ensures crop health and productivity. Fast and accurate results aid farmers in controlling diseases and ensuring healthy crops. This paper proposes the extensive residual network (ERN) model for generating high-resolution images and identifying diseases from low-resolution orange fruit images. The ERN model achieves significant results on the orange fruit greening disease dataset, obtaining high PSNR values 31.98, 32.785, and 34.853, SSIM values 0.8164, 0.9057, and 0.9164, and classification accuracies 99.24%, 98.71%, and 97.46% for super resolution factors of 2, 4, and 6, respectively.