Deep learning based vision for driverless vehicles in hazy environmental conditions

Technology and automotive companies have contributed ever increasing resources to the aim of developing self-driving or fully autonomous vehicles. While traditional robotics techniques provided initial successes in this area, machine learning and deep learning have become a critical technology for the further development in this area. Specifically, in the area of machine vision, there is a need for designing a robust object detection system with respect to negative environmental conditions such as haze. Therefore, I have created a novel training method for a deep learning based dehazing model that provides superior performance when compared to existing methods.