Pedestrian Statistics Based On KCF Tracking Algorithm

In recent years,with the continuous development of large cities,more and more people are pouring into big cities,which brings some pressure to the traffic of cities.By means of technical means,we can monitor the flow of people at the entrance,inform the subway service personnel in time,and take the necessary measures to realize the diversion of the personnel.At present,the technology of vehicle flow monitoring is developing rapidly.It also applies to all walks of life.The application of human flow detection and tracking technology can provide technical support for subway diversion,shopping mall control and traffic volume.This article is based on this,based on the integration channel and KCF tracking technology,the number of pedestrian statistics research work,the specific content includes the following.First discuss KCF tracking technology.KCF is a discriminant-based tracking method,which usually trains a target detector in the tracking process,uses the target detector to detect whether the predicted position of the next frame is a target,and then uses the new detection result to update the training set and then update the target detector.The positive and negative samples are acquired by using the circulant matrix around the target area,the ridge detector is used to train the target detector,and the diagonalizability of the circulant matrix in Fourier space is successfully used to transform the matrix operation into Hadamad product of vectors,Dot multiplication,greatly reducing the amount of computation,improve the speed of operation,the algorithm to meet real-time requirements.After comparing the KCF algorithm with other algorithms,the average precision and average FPS of the KCF algorithm are superior to other algorithms.In addition,the HOG feature is added to the KCF algorithm to make it more powerful improve.Then we design a human traffic statistics algorithm,based on the integral channel for people flow detection,based on KCF to achieve human flow tracking.The flow of the algorithm is as follows: Firstly,the first channel of pedestrian detection is made by integral channel;after the first frame is detected,the detected pedestrians are tracked,then the tracked pedestrian area is set to black,then the pedestrian detection is restarted,Pedestrian tracking initialization,the final completion of pedestrian count.At last,we use OpenCV technology to design the integral channel and KCF algorithm so as to realize the statistics of the number of people and achieve the purpose of fast and effective statistics of human traffic.