The proposed SRCNN model makes use of a three-layer convolutional network for super-resolution technique advancement, employing a 9 × 9 kernel for the feature extraction at a global scale, 3 × 3 for localized feature extraction, and 5 × 5 for the reconstruction of images at high resolution. It achieved a major boost in test data using both low- and high-resolution image pairs during training, delivering a PSNR of 29.02 dB and an SSIM of 0.869. PSNR represents the image reconstruction quality, while SSIM also captures the details and overall appearance. Results here indicate that SRCNN may be of great use for applications that require precise restoration of images —like medical imaging and satellite imagery—in which high-resolution standards are a good aspect of any analysis.