Design And Implementation Of Passenrgers Statistical System For Bus

Relevant data shows that most residents would prefer to take a cheap and convenient bus when they go out. But this transportation also has many problems,such as passengers waiting for a long time at some sites, some buses are too crowded, some buses of other lines running without passengers, etc. All these situations attributed to the bus company can’t receive the status of the running buses timely.The passengers statistical system for bus statistics for every sites the number of passengers getting on or off in specific time period, and give the manager feedback about the analysis results timely, which used to help bus company vehicle scheduling reasonably. This thesis expounds five aspects which include the passengers’ detection,tracking, counting, storage and display, the design and implementation of simulation software:For the passengers’ detection, this thesis implemented three algorithms, and then analyzed the experiment results on three aspects, which are the time of background generation, detection efficiency and accuracy. And we choosed the Haar-like classifier which trained by Adaboost algorithm, the method applied many group of head’characteristics to distinguish the passenger target, it’s not only achieve higher detection rate and lower false alarm rate, but also have clear target characteristics which not affected by carriages bumpy, it’s competent to do the job of detection on the bus.In the part of target tracking, this thesis analyzed the characteristics of running status of the bus and passengers moving, etc, thereby calculate the moving layout of passenger like speed dial which narrowed the search scope. Then changed the two-dimensional into a one-dimensional hash table, this way raised the search efficiency. Last seeking the optimal match in specific areas to treated them as one whole object in the time and space dimension, in other words, achieved the purpose of tracking.In the link of statistical passengers, first, this thesis utilized three-layer threshold to filter to acquire the "high quality" target which can participate in counting, and then according to the difference between the beginning and the end of the trajectory coordinates identified the direction, simultaneously counts the number of passengers getting on or off respectively. In order to ensure the accuracy of traffic statisticalresults, the system verify the validity of the results with the help of the number of total passengers on the bus when the bus stops before and after. In addition, this thesis briefly described the main database table design of the traffic statistics results.Finally, we have been carried out simulation to verify for the above algorithm. At first, this thesis simply mentioned the development environment of simulation software and main function modules, and then described the realization process of these modules in detail, meanwhile showed the design of critical codes in the process of system development. we tested the software based on some good test cases to validate the practical of simulation software. The test results show that, the simulation software which integrates a variety of high quality algorithm is discussed in this thesis have a better stability and higher operation rate and practical.