【作者】 陈剑锋；

【导师】 罗伟林；

【作者基本信息】 福州大学 ， 机械工程（专业学位）， 2020， 硕士

【摘要】 随着电子技术的不断发展,其应用范围日益扩大,相关的技术要求也越来越高,特别是在数据处理速度、实时性这两个方面。作为电子技术的一个重要应用,图像处理是一项数据量非常庞大的工作,在复杂的图像处理前先进行图像的预处理是非常必要的,设计合理的预处理不但能减少图像数据量,同时也能保证后续图像处理的质量。本文针对图像数据量大、噪点多、边缘轮廓模糊等关键技术问题,设计了一套完整的基于FPGA的实时图像预处理系统,主要工作内容如下:(1)通过对FPGA芯片内部结构的分析,基于其并行处理数据的能力和流水线的设计方式,确定了以FPGA作为控制核心的系统硬件平台,实现图像的实时采集、存储、算法处理、显示。(2)完成了系统的硬件搭建,包括OV5640摄像头、SDRAM控制器、VGA显示控制三大部分。设置摄像头工作模式,通过IIC协议完成摄像头配置,图像实时传输到FPGA中;优化SDRAM控制器的结构,采用乒乓读写操作,通过两个FIFO实现图像数据的无缝隙缓存,提高了存储器的存储性能;完成了FPGA与GM7123模块之间的接口设计,实现FPGA与VGA显示器的数据传输。(3)针对传统软件在图像预处理上的不足,结合系统对实时性的要求,在FPGA上实现了图像预处理的综合算法。采集到的图像经过灰度转换算法,降低了数据量;边缘提取前加入中值滤波算法,去除了噪点;边缘提取后,为了使边缘的连贯性和轮廓加强,加入了形态学滤波算法,保证后续图像处理的质量。(4)将各个算法模块整合到系统中去实现,完成一个完整的图像预处理系统。在Modelsim软件平台上,对系统的各个算法模块和硬件设备模块进行仿真验证,分析RTL视图和时序波形图。深入研究系统各部分的时序,主要是IIC时序、SDRAM时序和VGA时序的设计,保证设计的严谨性。根据系统资源消耗情况和图片处理效果,对系统做出合理的评价。本文设计的图像预处理系统已经在实际环境下进行了测试。输出显示的效果满足预期的设计要求,不仅满足实时性和处理速度的要求,而且实现了图像去噪、数据量小、边缘细节完整的图像预处理功能。更多还原

【Abstract】 With the continuous development of electronic technology,people have higher and higher requirements for image processing technology and applications,especially in terms of data processing speed and real-time performance.Image processing is a huge amount of data.It is necessary to pre-process the image before advanced image processing.This reduces the amount of image data and also ensures the quality of subsequent image processing.This paper designs a complete FPGA-based real-time image preprocessing system for key technical issues such as large amount of data in the early stage of image processing,more noise,and blurred edge contours.The main tasks are as follows:(1)This paper analyzes the internal structure of FPGA chip,based on its parallel data processing ability and pipeline design method to determine the system hardware platform with FPGA as the control core,and finally realizes real-time image acquisition,storage,algorithm processing,and display.(2)The hardware construction of the system is completed in this article,including the OV5640 camera,SDRAM controller,and VGA display control.Set the camera working mode: complete the camera configuration through the IIC protocol,and the image can be transmitted to the FPGA in real time;Optimize the structure of the SDRAM controller: using ping-pong read and write operations,two FIFOs realize seamless buffering of image data,and ultimately improve the storage performance of the memory.The interface design between FPGA and GM7123 module was completed,and finally the data transmission between FPGA and VGA display was realized.(3)Aiming at the shortcomings of traditional software in image preprocessing,the article combines the system’s requirements for real-time performance,and finally implements a comprehensive image preprocessing algorithm on FPGA.The collected image is subjected to a grayscale conversion algorithm to reduce the amount of data;a median filtering algorithm is added before edge extraction to remove noise;After edge extraction,in order to enhance the consistency and contour of edges,a morphological filtering algorithm is added to ensure the quality of subsequent image processing.(4)It is realized by integrating various algorithm modules into the system to complete a complete image pre-processing system.On the Modelsim software platform,simulation verification is performed on each algorithm module and hardware equipment module of the system,and the RTL view and timing waveform diagram are analyzed.The in-depth study of the timing of each part of the system is mainly the design of IIC timing,SDRAM timing and VGA timing to ensure the rigor of the design.By using the system resource consumption and picture processing effects,a reasonable evaluation of the system is made.The image pre-processing system designed in this paper has been debugged and tested in a real environment.The system output display effect meets the expected design requirements.It not only meets the requirements of real-time performance and processing speed,but also implements image pre-processing functions of image denoising,small amount of data,and complete edge details.更多还原