Vision-Based Front and Rear Surround Understanding Using Embedded Processors

Vision-based driver assistance systems involve a range of data-intensive operations, which pose challenges in implementing them as robust and real-time systems on resource constrained embedded computing platforms. In order to achieve both high accuracy and real-time performance, the constituent algorithms need to be designed and optimized such that they lend well for embedded realization. In this paper, we present a novel two-camera-based embedded driver assistance framework that analyzes the dynamics of vehicles in the front and rear surround views of the host vehicle (ego-vehicle). In order to do this, we propose a set of integrated techniques that combine contextual cues and lane information to detect vehicles that pose high threat to the ego-vehicle. The threat analysis is then used for generating a safe maneuver zone by the proposed system, which is implemented by using two Snapdragon 810 embedded CPUs. A detailed performance evaluation and tradeoff analysis is presented using a novel multiperspective dataset (DualCam Dataset) that is released as part of this paper. In terms of accuracy, the detailed evaluations show high robustness with true positive rates greater than 95% with less than 6% false alarm rate. The proposed embedded system operates at real-time frame rates in our testbed under real-world highway driving conditions. The proposed framework was presented as a live demonstration at the 2016 Consumer Electronics Show.