| As a kind of label,2D code is widely used in industrial inspection field due to its large storage capacity,strong reliability of decoding information and fast recognition speed.In the process of dynamic product inspection,relative motion will be generated between the image acquisition device and the product,resulting in motion blurring of the QR code,which increases the difficulty of system identification and reduces the real-time performance,so it is especially important to read the QR code information efficiently.This thesis addresses the problems of large size,high power consumption and poor real-time performance of QR code recognition through the traditional PC platform,and designs a QR code acquisition and recognition system based on ZYNQ embedded platform in a high-speed environment,and carries out research in several aspects,including system architecture construction,hardware acceleration of pre-processing algorithm,optimization of motion fuzzy recovery algorithm and porting of embedded operating system,respectively,as follows:1)Build the ZYNQ platform 2D code high-speed acquisition and display system,and complete the digital logic optimization design between each module.The OV5640 COMS image sensor acquisition unit,image cache unit and HDMI interface display unit are designed using the modular design method,and the data communication between ARM and FPGA is realized through the high-performance AXI bus,which significantly reduces the transmission delay of the whole acquisition and display system and improves the real-time performance of the system.2)The FPGA-based hardware acceleration module for 2D image pre-processing and target detection is designed.For the problems of noise interference and low contrast in the acquisition process,the median filtering and histogram equalization modules are designed using the idea of FPGA pipeline parallel execution,which solves the problems of noise interference and low contrast in the acquisition process.For the problem of difficult positioning of 2D codes due to high speed motion,we designed a method of image positioning based on connected domain marker by using the read/write feature of FPGA part RAM,which improves the speed of 2D code positioning and saves system hardware resources.3)The motion blur 2D code image recovery algorithm is implemented based on ZYNQ platform.Firstly,the degradation process of two-dimensional code in high-speed environment is analyzed,and the motion blur degradation model is constructed.Secondly,in view of the problem that it is difficult to estimate the motion blur kernel function in the spatial domain,a multi-channel FFT module is designed in the FPGA part of ZYNQ to improve the efficiency of spatial domain transformation.Finally,the traditional Wiener filter image restoration algorithm is improved,and the improved algorithm improves the peak signal-to-noise ratio(PSNR)of the image after the restoration of the traditional algorithm by about 4d B,so that the image quality after restoration is significantly improved4)The embedded Linux operating system is ported and the application software is designed.In order to realize the control acquisition of 2D code image,motion blur recovery and display multi-task parallel execution,the Linux operating system was ported in the ARM part of ZYNQ to improve the controllability and flexibility of the system.The configuration and mirroring of Open CV and Qt libraries were completed,and the recognition of 2D codes and the display of human-machine interface were realized.This thesis tests the functions of each module of the system by building a simulation test platform,verifies the hardware acceleration and image restoration effect of the system,and analyzes the resource occupancy rate and power consumption of the system,the results show that the system can achieve stable identification of two-dimensional code at 0.3m/s operating speed,and the power consumption of the system is 1.98 W,which meets the requirements of system design,realizes the needs of low power consumption,low cost and convenience,and the system has great engineering application prospects. |
