A New Velocity Estimation Based on Image Intensity Curve
Zhengxi Li · Jisuanji fangzhen · 2009
A real-time vehicle speed estimation algorithm is proposed. It uses image intensity curves of each lane to estimate the mean speed. First, the lane intensity curve is generated by divided the image into different lanes. Then the relation between the previous frame and the next frame is applied to find the best match of the lane intensity curve by using minimize the math error. Finally the mean speed of the lane is computed by using the weighted mean of all point speeds. Our algorithm is effective for high resolution image using its less complexity. The vehicle speed estimation results using high resolution image in a highway have shown our approach’s effectiveness.