| Moving objects detection and tracking in video sequences is a task related to videoimage processing, computer vision, pattern recognition, artificial intelligence andother areas. It is an important research topic all the time in recent several yearsbecause of its widely application in business and military fields. With theestablishment of the tobacco enterprise and the integration of tobacco brand, thecompetition among the enterprises turns from a single brand competition into a brandand cost competition, so improving the level of cigarette manufacturing becomes theultimate goal. In this paper, the moving target detecting and tracking technology isused on the detection of tobacco transfer speed and it is an attempt to use thetechnology based on the temporal difference and block matching to calculate thespeed of tobacco.First of all, use the camera to collect the video sequence images of tobacco in thepipe, and then extract the useful information by processing these images. Secondly,compare the detecting and tracking methods according to the tobacco’s motion state.The first design scheme--use the full search block matching method to estimate thetobacco’s average motion vector of every frame; the second scheme--use the methodwhich combines frame difference with full search block matching to estimate thetobacco’s average motion vector of every frame. Lastly, use VFW to realize videocapture under the MFC framework of the Visual C++6.0, edit and debug thealgorithm routine, display the data and path chart on the interface.The experimental results show that the two schemes of this design can bothdetect the velocity of tobacco in pipe real-time on-line. The dates get from the secondscheme, by contrast, update faster and more sensitive, can achieve the requirement. Sothe design of research on the speed detection of tobacco based on temporal differenceand image block matching has the advantage of simple operation and low-cost etc. Ithas a good prospect. |
PDF Full Text Request