为了从广泛使用的智能卡付费系统获取公交客流信息,提出了一种利用公交调度信息和智能卡刷卡信息推断乘客上车站点的方法.对同一辆车的连续2次刷卡进行朴素贝叶斯分类,区分是否是在同一个站刷卡;利用极大似然估计、动态规划和二次规划方法估计出各路段的行程时间;运用坐标下降法从不准确的初始参数出发,交替估计行程时间和行程时间的参数,从而推断出每次刷卡的上车站点.实验结果验证了新方法的正确性和有效性,证明了该方法误差较小,收敛较快. In order to obtain bus passenger information from the widely used payment system of smart card, a method to infer passengers’ boarding stations using bus dispatching information and smart card swipe information is proposed.A simple Bayesian classification of two consecutive credit cards of the same car is made, Whether it is in the same station credit card; using maximum likelihood estimation, dynamic programming and quadratic programming method to estimate the travel time of each section; using coordinate descent method from inaccurate initial parameters, alternately estimated travel time and travel time Parameters to infer the boarding station for each credit card.The experimental results verify the correctness and validity of the new method and prove that the proposed method has less error and faster convergence.