| Popis: |
In order to reduce traffic accidents involving animals, which is a major concern in worldwide traffic, Autoliv has developed a state-of-the-art vehicle mounted night vision animal detection system. The system is currently used by Audi, BMW and Daimler. The main contributions of this paper include: world's first vehicular animal detection system to reach the customer market, an efficient classification approach based on a cascade boosting concept which is robust to occlusion, pose and scale variations, a large database of thousands of hours of far infrared (FIR) video data recorded worldwide including several hundred thousand example images of animals in traffic situations, a tracking approach to handle animal movement and estimate animal states, a validation approach to efficiently reduce the number of false detections and human-machine-interface (HMI) and warning concepts to highlight animals at risk of collision. The presented system detects animals up to 200 meters away from the car while generating very few false warnings. For animals that are considered a potential danger, advanced HMIs such as marking lights which actively illuminates the animals are applied, giving the driver the quick and accurate information he or she requires. The Autoliv night vision animal detection system is complementary to currently used methods for preventing accidents with animals. By using it, the driver is given all opportunities to react to dangerous situations and to avoid potential accidents. |
