Road region detection is an essential perception functionality to the mobility of autonomous vehicles. It can also help human drivers to drive safely in challenging weather conditions such as low visibility, snow storms and heavy rain. The majority of current road region detection techniques use camera and lidar sensors, but they struggle in inclement weather conditions due to the limitation of feature extractions in these conditions. In this talk, I am going to highlight how we use MMwave radar to provide accurate road region detection in challenging conditions. I will also describe a novel approach to overcome the sparse resolution of mmwave-radars to provide better results in image segmentation tasks.