基于FPGA的实时监控系统和边缘检测的研究与实现

发布时间：2018-07-09 15:27

本文选题：摄像头 + 门阵列　；参考：《天津理工大学》2017年硕士论文

【摘要】：随着工业自动化的不断发展,基于视觉的监控系统和图像处理系统的需求在持续增长,在公共交通、安全监控、环境监控、工业检测、卫星与实验等领域有着重要的应用。而在目前的视频监控系统中应用FPGA处理的并不是很多,由于嵌入式系统中基于FPGA、DSP等处理芯片的能力越来越强,尤其是FPGA的可并行处理优势,一些像素级的复杂算法在FPGA上可以轻而易举的实时完成,如视频的编解码、边缘提取、边缘检测等,所以FPGA势必成为今后视频监控领域的一面旗帜。本文深入分析了现场可编程门阵列(FPGA)的基本结构和工作方式,提出了利用FPGA并行处理优势完成实时监控系统的搭建和实现方案,并实时完成图像的灰度化、边缘检测、腐蚀膨胀算法的实现,得到较好的边缘检测效果。该系统采用了ALTERA公司的Cyclone IV系列作为核心处理芯片来驱动OV7725摄像头,对采集的数据进行格式转换、灰度化、边缘检测、膨胀腐蚀等算法处理,最后驱动VGA实时显示。与传统系统相比较,此系统具有可重构性、实时性好、设计灵活、处理速度快等优势。系统主要分为图像采集模块、预处理模块、边缘检测模块、跨时钟域和图像缓存模块、VGA显示模块等,此系统利用verilog HDL语言对各个模块编程实现,在Quartus II平台上完成测试和板级实现,尤其对OV7725摄像头输出的视频流在不同时钟域实现跨时钟域的数据交互、边缘检测和膨胀腐蚀算法的FPGA实现深入分析,利用modelsim和signal Tap II软件进行了仿真和板级调试。研究结果表明,在FPGA硬件平台上能够高效的完成实时监控以及边缘检测等复杂算法的实时实现,利用图像灰度化、腐蚀膨胀算法使得边缘检测效果更平滑、清晰,满足今后复杂高精度的监控安全、航空航天领域的图像处理实时性要求,对以后目标识别、检测、导航制导等复杂算法的研究具有较高的理论和应用价值。
[Abstract]:With the continuous development of industrial automation, the demand of visual monitoring system and image processing system is increasing continuously. It has important applications in the fields of public transportation, safety monitoring, environmental monitoring, industrial detection, satellite and experiment and so on. However, there are not many applications of FPGA in the current video surveillance system, because the ability of processing chips based on FPGA DSP is becoming stronger and stronger in embedded system, especially the advantages of FPGA in parallel processing. Some complex algorithms at pixel level can be easily implemented in real time on FPGA, such as video coding and decoding, edge detection, so FPGA will become a banner in the field of video surveillance in the future. In this paper, the basic structure and working mode of Field Programmable Gate Array (FPGA) are analyzed in depth, and the scheme of real-time monitoring system based on FPGA parallel processing is put forward, and the image grayscale and edge detection are realized in real time. A better edge detection effect is obtained by the implementation of the corrosion expansion algorithm. The system uses Cyclone IV series of Altera company as the core processing chip to drive the OV7725 camera. The collected data are processed by format conversion, grayscale, edge detection, expansion corrosion and so on. Finally, the real-time display of VGA is driven. Compared with the traditional system, this system has the advantages of reconfigurable, real-time, flexible design and fast processing speed. The system is mainly divided into image acquisition module, preprocessing module, edge detection module, cross-clock domain and image buffer module, VGA-display module, etc. Test and board level implementation is completed on Quartus II platform. Especially, the FPGA implementation of OV7725 video stream in different clock domain, edge detection and expansion corrosion algorithm is deeply analyzed. Modelsim and signal Tap II software are used for simulation and board level debugging. The research results show that real-time monitoring and edge detection can be realized efficiently on FPGA hardware platform. The edge detection effect is smoother and clearer by using image grayscale and corrosion expansion algorithm. It can meet the requirements of complex and high-precision monitoring and control and real-time image processing in the field of aerospace. It has high theoretical and practical value for the research of complex algorithms such as target recognition, detection, navigation guidance and so on.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】