• Smitha K
• Rejeena .K.A

Aerial image restoration plays a major role in applications such as urban planning,environmental monitoring, and disaster management, where high-quality imagery is crucial for accurate decision-making. However, aerial images often suffer from challenges like noise, poor contrast, color distortion, and atmospheric interference, which hinder effective analysis. Traditional restoration techniques struggle to address these multifaceted issues, particularly in complex scenarios. This study proposes an advanced approach integrating digital image processing techniques with deep learning, specifically using CycleGAN (Cycle-Consistent Generative Adversarial Networks), to restore aerial images. The methodology begins with a comprehensive image restoration pipeline, including color stabilization, saturation enrichment, and contrast improvement, followed by deep learning (DL) enhancement through CycleGAN. The restoration process is carried out on the USC-SIPI Image Database, focusing on aerial imagery. Key image characteristics such as saturation, contrast, and brightness are adjusted to enhance the clarity and usability of the images. The CycleGAN model, operating on unpaired datasets, enables effective translation from low-quality to high-quality images, preserving essential features and improving visual appeal. Evaluation metrics, including Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Mean Absolute Error (MAE), were employed to assess the performance. The proposed model attained a PSNR of 38.34, an SSIM of 0.9806, and an MAE of 0.005, demonstrating superior restoration quality compared to existing methods. These results highlight the model's capability to improve image quality while conserving critical facts. The Cycle GAN-based approach provides a powerful and adaptable solution for real-world aerial image restoration, offering significant improvements over traditional techniques and making it a valuable tool for applications requiring high-fidelity image analysis.

Aerial Image Restoration, Deep Learning, Digital Image Processing, ColorStabilization, Contrast Enhancement, Image Quality Improvement.

"Aero Vision Net: A Deep Learning-Powered Framework for Aerial Image Restoration with Advanced Digital Enhancement Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 12, page no.d10-d29, December-2024, Available :http://www.jetir.org/papers/JETIR2412302.pdf

"Aero Vision Net: A Deep Learning-Powered Framework for Aerial Image Restoration with Advanced Digital Enhancement Techniques", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 12, page no. ppd10-d29, December-2024, Available at : http://www.jetir.org/papers/JETIR2412302.pdf