|
The rapid development of intelligent transportation system to accelerate the pace of transportation into the information age , passenger counting system is an important part of traffic information , it is an urgent problem , get the full attention of relevant departments . Traditional statistical methods , there are many drawbacks , therefore , need to explore a low-cost , real-time , feasible method . Infrared sensor - based counting method has its unique advantages , will be one of the best choices . Dynamic time warping algorithm (Dynamic Time Warping, DTW) universal applications in speech signal processing to achieve high accuracy , and in isolated word recognition applications . Due to the acquisition of the passengers to get on and off infrared signals and voice signals are very similar , so I try to the introduction of the algorithm on and off automatic passenger counting system . In this paper, a clustering algorithm to select a highly adaptive signal as a reference template , by extract the analyte signal the endpoint detected from the sampled signal sequence , and with the reference template matching , select match the best of the best results , to statistics the number of passengers on and off . Dynamic time warping algorithm is highly dependent on the results of the endpoint detection , in order to get a high statistical accuracy , good must signal endpoint detection . In this paper, on the basis of careful study of the characteristics of signal endpoint endpoint detection methods applicable in this topic , and DTW algorithm matching function Search improved dynamic time warping algorithm to get a good application in the field . Active and passive infrared sensor fusion , VC environment endpoint detection , DTW algorithm implemented in software . Intuitive observation endpoint detection the DTW algorithm design validity , this paper carried out off-line simulation . The simulation shows that the signal endpoints and identify the number of on and off , the statistical results showed that the accuracy rate of up to 97% can be well defined . Counting system because of its low cost , safety , and has the advantages of non-contact infrared sensor , in shopping malls , stadiums , and many other occasions can be used , so the system has good practical value and broad market prospects .
|