Enhancing of Features for Road Crack Image Using EEcGANs

Continuously monitoring roads to examine their defects is a difficult task for the human element due to the increasing length of roads. Therefore, several types of digital imaging tools are used, for example (drones, surveillance cameras fixed on poles and cameras fixed on a moving car). These imaging tools automatically take pictures under various weather conditions and the blurriness generated by cameras mounted on moving cars. This paper proposes a method to enhance road surface images captured under different conditions, where the image will be improved in the frequency domain and spatial domain and then added to the enhancement with the Enlighten of Conditional Generative Adversarial Networks (EEcGANs) to obtain a generated and enhanced image. The metrics Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Difference in Variance (DIV), Correlation Coefficient (CC) and Universal Quality Index (UIQ) are used to evaluate the performance of the proposed method, achieving the values 29.4019, 0.8996, -0.12151, 972758 and 0.970491, respectively.