Driver assistant systems and functions are becoming a general requirement in modern cars, but many of these functionalities still pose significant challenges in image processing and machine learning. Although algorithmic solutions with sufficient accuracy exist for many tasks, most commonly these methods require complex and expensive infrastructure (both computational and sensory) and have high power consumption. In this paper we present the ReCo(exp TM) - Pro, a cellular vision system with advanced driving assistance functionalities. The system is capable of detecting and tracking various objects: vehicles, pedestrians, traffic signs and traffic lights with 60 frames per second in HD resolution and with 5 watts power consumption.
展开▼
