Font Size: a A A

Research On Video Analysis For Actual Loading Rate Monitoring Of Long Distance Coach

Posted on:2014-06-13Degree:MasterType:Thesis
Country:ChinaCandidate:M J WangFull Text:PDF
GTID:2298330467951739Subject:Transportation planning and management
Abstract/Summary:PDF Full Text Request
Long-distance passenger transport industry remains in transportation industry market disorder, generalized illegal business, frequent road passenger traffic accident and poor quality of service since it existed in our country. This shows that monitoring the long-distance passenger coach is very necessary, and to monitor the actual loading rate index of the coach can achieve multiple monitoring purposes, In view of the actual load rate monitoring problem, We design a system which is suitable for monitoring the actual loading rate of the long distance passenger coach.First, we compared a variety of monitoring technology, such as pressure sensing technology, Infra-red monitoring, photoelectric technology, video analysis, etc. And we selected video analysis technology as the basis of the research technique. So an actual loading rate monitoring system based on video analysis was proposed.Secondly, there is no a general algorithm suit for all situations now owning to the application of video analysis in the passenger flow counting is still a challenging research subject. Thus we focused on video analysis techniques which including target detection, target tracking, target recognition. In terms of target detection, we put forward a gridding moving object detection algorithm based on background subtraction and frame differencing, and confirmed that the algorithm is suitable for target detection in different monitoring scenario. In terms of target tracking, a target tracking algorithms based on CamShift and another one which based on centroid point were applied to the passengers tracking problem. By comparing the two methods above, we found out the tracking algorithm based on centroid point plays better results. In terms of passenger counting, the paper proposed three passenger counting methods respectively based on the single inspection, the onset and offset of target tracking and pixels’statistics, those passenger counting methods are proved practicable through experiment. In the passenger counting based on pixel statistics, an algorithm based on weighted step peak finding is raised, which can effectively identify the signal wave carrying a lot of burrs, but with obvious peaks and troughs.Finally, on the basis of the above theoretical research, we designed a simple passenger actual loading rate monitoring system constituted by passengers counting system, overload alarm system and settlement report system, and Visual C++6.0was used as the system software platform. Using the self-collected video data to test this system, the result shows that the system is able to achieve the expected function.
Keywords/Search Tags:Actual loading rate, Automatic passenger counting, passenger monitoring, Target detection, Intelligent identification
PDF Full Text Request
Related items