The pedestrian detecting and counting system based on automatic method of CCD calibration
Yuejin Wang, Sumei Guo, Hongwei Huang · 2017
A pedestrian detection and counting method is proposed in this paper, which can effectively detect pedestrians when they appear in the scanning area. This method does not require manual intervention to achieve camera calibration, the detection process can be carried out rapidly and accurately, and the use of machine learning methods Haar+Adaboost classifier are trained to determine pedestrians. At the same time, the hybrid Gauss background model is adopted to separate the foreground from the multi frame images, and then the local minimum matrix method is used to further lock the pedestrian change range. The method can real-time calculate and calculate all the pedestrians in the video window, and automatically update the camera parameters at the same time by iterative method. Experimental results show that this method can realize real-time pedestrian detection and counting with low false detection rate, so it is feasible and effective.