License Plate Recognition Technology Based On The Hardware Acceleration Technology On The ZYNQ

With the improvement of people's living standard and the development of urbanization,the increasingly serious traffic congestion situation in the city has gradually become an urgent problem to be solved.As an advanced and efficient means of traffic management of our era,license plate recognition system is one of the important research direction.However,this technology based on the PC platform has so many disadvantages such as high cost,large volume,high power consumption that can't be widely used.With the rapid development of embedded development technology and FPGA,many embedded products become more and more popular,which are easy to carry and operate,and lower cost.So taking a study on the license plate recognition technology based on the FPGA has great significance and broad prospects.It studies some classic algorithms of license plate recognition and development techniques on the ZYNQ platform.First,according to the characteristics of the license plate area,the edge detection and morphological processing of the license plate are carried out by the improvement of the Soble operator and the way of dynamically setting the structural elements,and the license plate area is extracted according to the geometric characteristics of the license plate.And then,for the tilted license plates,using the improved Hough transform that reducing the number of pixels to be processed to correct them.Secondly,by analyzing the drawbacks of the traditional projection method,we choose to calculate the ordinate range of each pixel column,and then draw the range curve,segment the characters by finding out the smallest seven values in the range array.Finally,it analyzes the three shortcomings of the traditional template matching.And based on the mathematical significance of the correlation coefficient in the mathematical statistics,the standardized similarity function is constructed to measure the similarity between the template character and the character to be recognized to recognize the license plate characters.On the basis of the research of the algorithm,a license plate recognition technology based on the ZYBO in the Zynq-7000 series is proposed.The platform consists of programmable logic(PL)and processing system(PS).The PL part can realize the control logic of each module of the system such as the control of the DMA transmission of image data and the hardware acceleration by the HLS;the job for the PS part is running the Linux on the ARM Cortex-A9 of the ZYBO and carring out the subsequent processing for the image from the part PL,and then control the IP core of license plate preprocessing.Finally,the experiments for the license plate recognition technology based on ZYNQ platform is carried out.One thousand license plate images of normal and complicated environment are used in the experiments.The results show that the license plate recognition rate under normal conditions,and complex environment are fairly high.Finally,comparing the time consuming of the image preprocessing part between the hardware and software scheme and that of the whole project.The results show that the overall time consuming in hardware scheme reduced by more than two seconds.The results show that the ZYNQ platform has the advantage of high performance and low power consumption in the field of image processing...