Infrared and visible image fusion aims to extract essential information and synthesize a single fused image containing both the salient target and texture details. Recent fusion techniques, including deep learning methods like auto-encoders, convolutional neural networks, and generative adversarial networks, have made significant strides but still face challenges such as inadequate texture and thermal detail capture, noise, and contamination. To address these problems, we propose a novel GAN-based fusion method that integrates an all-in-one dehazing network (AODNet) with a conditional generative network featuring double discriminators. We perform edge loss calculation on the generator network. This method preserves pixel and structural details, enhances edge sharpness, and effectively eliminates haze and blurry occlusion, resulting in superior image quality even under low-light and noisy conditions. Our contributions include improved accuracy in image fusion, better edge and boundary definition, and enhanced clarity of distant objects.