Automotive Vision Camera System based on Neuromorphic Processor

This book describes automotive vision camera systems to provide road scene information necessary for autonomous driving vehicles to plan driving routes and establish driving rules. For this purpose, a lane detection system and a traffic sign recognition (TSR) system, which can be implemented by an image sensor alone, were designed based on a neuromorphic processor and evaluated through the experiments. The vision camera system for autonomous driving vehicles must ensure the stability of the system for the safety of the driver and also the flexibility of the system for high recognition accuracy. In this book, in order to achieve the stability and flexibility of the automotive system, the neuromorphic network input was diversified into pixel values or statistical pattern vectors of the original image according to the pattern characteristics of the lanes and the traffic signs.