Development Of Bus Passenger Flow Detection System Based On FGD Technology

As the main transportation of city residents travel,bus provide power as if the artery for the city's normal operation.However,due to the development of economy,urban traffic situation is becoming more and more complex,There are obvious shortcomings in the operation mode of the old bus system,especially the number or time of flights and so on in the peak passenger flow.The bus has many ways to ride and difficulty in real-name system,so it is not easy to count bus traffic.From ancient times,image is the main mean to learn knowledge,grow awareness and access information for humanity,as early as the text did not formed,the image plays an important role in recording and disseminating information.With the development of the times and the continuous progress of science and technology,computer network is increasingly developed,image processing technology is improving in our life,using computer image processing technology to help people complete the integration and processing of a variety of complex information has become more simple,using image processing technical to count bus traffic is one of them.In recent years,there have been many ways to detect passenger flow in various public places.The method based on infrared is very widely recognized and applied.However,the infrared method is not very fit for the complicated environment of the bus,such as repeated through the detection area may cause repeated counting the case,and we need to add infrared equipment on the bus;However gravity-based statistical technology is also difficult to achieve dense passenger flow volume statistics,and testing equipment is easy to wear or even damage.At present,the statistical method based on video sequence has gradually become the main means,using frame difference method,background difference method,optical flow method and other means to lock the target area to detect the moving target,and then through the analysis of differential data to count passenger flow volume.But the frame difference method and other methods exist inherent defects,such as the frame difference method in detecting position is sometimes not accurate enough,for the complex passenger situation in bus,we can not apply these methods mechanically.In thispaper,an improved algorithm is proposed according to the disadvantages off difference method,which is to convert the difference value after this method into a one-dimensional characteristic array.Through the non-linear filtering smoothing feature,the effective mutation point is retained while noise interference is excluded.In addition,the detection area is specified in the video,and for the characteristic array in different area,we use the normalization process and the continuous numerical detection method to calculate the number of targets passing through the detection area.The method is as follows: Firstly,the region of interest is generated by the background difference method,then we can obtain the difference sequence by differentiating the Y component in the region of interest,after it,we quantize the difference sequence and perform median filtering and adaptive threshold,finally,we have the detection and decision for the waveform.In this paper,the algorithm is applied to the Hi-3515 platform based on Linux system,a passenger flow statistics system is built and verified experimentally,besides,a dense passenger flow counting system is achieved.Experiments show that this algorithm can significantly reduce the calculation cost of frame difference method,realize real-time detection and improve the accuracy rate of dense population flow counting.