<p>Rain and fog are common weather conditions that can significantly degrade the quality of outdoor images. In this research, a novel method, LDD-Net is proposed for defogging and deraining a single image using a lightweight CNN model. The technique leverages the capabilities of CNNs to remove rain streaks and fog from images efficiently. To overcome the limitations of existing methodologies, two CNN networks are designed with identical architectures, with the output layer being the only difference. They are designed to remove rain and fog from the input image and to extract a clear image by enhancing the visibility of the input image. These networks are trained independently, and then their results are ensembled to improve the output. To evaluate the effectiveness of the approach, qualitative and quantitative evaluations are conducted. The effectiveness of the method in enhancing image clarity and removing rain streaks and fog is demonstrated by visual comparisons. Quantitative metrics, such as the PSNR and the SSIM, further validate the superiority of the method over existing techniques. Additionally, the proposed method achieves fast inference, enabling real-time application in various scenarios.</p>