Tightly-coupled LIDAR and computer vision integration for vehicle detection

摘要: In many driver assistance systems and autonomous driving applications, both LIDAR computer vision (CV) sensors are often used to detect vehicles. provides excellent range information different objects. However, it is difficult recognize these objects as vehicles from alone. On the other hand, imagery allows for better recognition, but does not provide high-resolution information. this paper, a tightly-coupled LIDAR/CV integrated system proposed vehicle detection. This sensing mounted on front of test vehicle, facing forward. The sensor estimates possible positions. then transformed into image coordinates. Different Regions Interest (ROIs) in defined based object hypotheses. An Adaboost classifier utilized ROIs. A error correction approach choose an optimal position detected vehicle. Finally, vehicle's dimensions derived data. main contribution paper that complementary advantages two utilized. scanning data applied correction. And output distance dimension Experimental results presented illustrate reliable. It can be applications such traffic surveillance roadway navigation tasks.