Design Of Underlying Algorithm Of Embedded Smart Camera Based On IP

With the increasing maturity of intelligent driving technology,intelligent driving cars have gradually entered the public's vision and become a hot spot in the current social.The development of smart cameras provides a new solution for the development of intelligent driving vehicles to obtain information about the surrounding environment.With the development of embedded,the traditional smart camera based on PC because of its big size,poor integration,long hardware upgrade cycle,upgrade maintenance difficulties,has been replaced by the embedded smart camera.Based on embedded technology,an embedded smart camera system based on hardware IP core is designed for intelligent driving.In this paper,I choose the Zed-board based on the ZYNQ SOC as the hardware platform of the whole system,the system is divided into three parts according to the requirements: the underlying hardware design,embedded system transplantation and man-machine interface design.On the hardware design,the realization of the function of the system is modular.The smart camera was divided into image acquisition,image processing and display unit.Image acquisition are the OV series and USB dual channel image acquisition.Image processing unit,the hardware IP of image processing algorithms was designed by VIVADO HLS,it transformed the high-level language code to the RTL code and encapsulated into image processing IP core.The images display by HDMI.it's mainly to transplant embedded LINUX operating system and QT graphical interface library and open source computer vision library,in the human-computer interaction interface using QT and Open CV realized the intelligence of traffic signs detection and recognition.After testing,detection and identification accuracy is above 85%,and the real-time processing rate is 8 frames/sec.