Roadway Image Preprocessing for Deep Learning-Based Driving Scene Understanding
Kyeong-Sik Park, Sung-Su Jang, Hyeok-June Jeong, Young-Guk Ha · 2019
In this paper, a road lane is detected as a reference point in a road driving image, and the objects that are likely to affect the driving situation are detected depending on a reference point, and then the remaining objects are excluded from the road driving situation recognition. In addition, we propose a method of mapping to Bird's Eye View as a normalization method to uniformly express the positions of objects irregularly appearing on each road driving image based on these reference points. As a result, the accuracy of 97.66% was obtained as a result of the experiment on the road driving situation recognition by the proposed method.