Study on Threshold Segmentation Algorithm in Video Vehicle Measurement
Tianmin Deng · Journal of Highway and Transportation Research and Development · 2009
In order to improve the precision of parameters of traffic flow in the video vehicle measurement,the threshold segmentation algorithm in the video detection was researched.The fractal dimension method,2-mode method,iterative method and Ostu method were analyzed for getting image threshold to make the vehicle separated with the background.The results show that(1) the fractal dimension method needs long time to deal with the image;(2) the iterative method has complex programming procedure and needs long time dealing the image;(3) the 2-mode method can be used for process of certain image with special requirement but it is unfit for the process of video vehicle image;(4) although the two-value image that dealt by Ostu method has black pixels in the vehicle,the separation of the vehicle from the background is better,so the Ostu method was selected as the threshold segmentation algorithm.