Vehicle detection from onboard camera using patch decided vanishing point
Zihao Wang, Weidong Qu, Sei‐ichiro Kamata · 2017
In this paper, a vision-based vehicle detection using vanishing point is proposed, which could detect vehicles with different orientation using different models, and increases the accuracy. With the input image from the on-board camera, the method first divides it into small square patches. Then two approaches are used for grey image patches and color image patches which help to find out useful patches for vanishing point vote process. After that, the method uses selective search to generate candidates that might be vehicles. And with the vanishing point, orientation and scale of candidates are obtained. According to these, they are put into corresponding models which are trained offline. A new data set included orientation is also created for vehicles and non-vehicles which are used to do the experiment. The result shows that the method works and improves the accuracy.