The Research And Application Of The Key Technologies For Traffic Flow Video Detection In Parking Lot

In the current situation of rapid economic growth,with increasing car ownership,more efficient use of existing parking spaces has been supposed to meet the needs of owners,and not only do the current parking lot detection equipment and display devices have high cost of hardware and software resources consuming,but also information communication between devices is closely related with the transmission distance and the transmission rate.Based on the above background,this paper studies the key technologies of traffic video detection,and applies it to the parking lot.The system has the advantages of portable hardware,low cost,intuitive and transparent parking information,convenient equipment maintenance and so on.Most of the traditional video detection algorithms use computer platform for data processing,and the system hardware platform adopts TMS320DM642 chip,with camera as data acquisition device,data encoding and decoding chip to digital analog conversion,extended SDRAM for video data storage and extended FLASH to store the algorithm program,finally through the GPIO port transmission signal to the external LED display system to complete the traffic monitoring of parking lot and real-time display of parking spaces.The system uses the traditional background difference method and threshold method to detect moving vehicles in the image,and the traditional backgroundupdating algorithm is improved,which smooths noise in the background and at the same timethe real-time of the system is guaranteed.In addition,to seek out the position and height about virtual detection zone,we analyzed experiment by establishing mathematical model,then obtained the updated speed of background model.Meanwhile,the adjacent frame two value classification algorithms is proposed to judge the direction of vehicle running,which bases on the law of the change of pixels to identify the direction of vehicle,the algorithm with relatively small amount of computation would meet the real-time requirements of the system.Finally,the vehicle is identified and counted by threshold segmentation to ensure the accuracy of the system.At the end of this paper,we simulate the car test and field test program,the results show that the system can meet the requirements of real-time monitoring and counting of vehicle flowin parking lot.