Challenging weather conditions such as fog pose a big problem in vehicle detection because it hinders easy recognition and affects the quality of the image even more. In this work, vehicle detection is investigated with an enhancement of accuracy for poor lighting conditions through the employment of YOLO deep learning algorithm coupled with Retinex image enhancement technique. YOLO is an advanced object detection algorithm that is fast and very precise but is not sensitive to low light areas. Likely to improve the image visibility and detection reliability by mimicking mechanisms of a human eye, the Retinex algorithm is capable to mitigate the negative impact of fog. The overall idea presented in this paper is based on the pipelined approach, where Retinex algorithm will be employed initially to improve the input images to be fed to the YOLO vehicle detection system. The effectiveness of this approach is tested using a set of foggy weather images revealing much higher level of detection performance and precision as compared to YOLO alone.