高速智能图像采集系统的设计

发布时间：2019-04-09 16:42

【摘要】：高速智能图像系统是兼具高速图像采集、智能化的图像处理和高速率的数据传输能力的嵌入式系统设备。随着信息产业的不断发展,智能图像采集系统日益趋向小型化、智能化、网络化和高速化。如何使更具智能化、低成本、多功能的图像设备应用于人们的生产生活已日益成为研究的焦点。本课题结合FPGA嵌入式图像开发技术和USB3.0接口的超高速数据传输技术,从智能化和高速化两个方面出发,尝试设计一款针对特定应用的具有低成本、低功耗和较好性能的高速智能图像系统。在高速性方面,利用Cypress公司EZ-USB FX3 USB3.0开发套件,设计针对1080P视频流传输的UVC接口;在智能性方面,以当前较为成熟的人脸肤色检测理论为基础,在FPGA实现视频图像的人脸片上实时检测。课题完成的工作有:首先,在MATLAB平台上验证基于肤色的人脸检测算法的可行性;其次,完成了FPGA端彩条图像的IP核设计和1080P图像传输的USB3.0接口设计,包括设计乒乓缓冲结构,减少高速数据传输出现的延迟;最后,完成肤色算法模型的硬件实现,通过搭载低速摄像头OV7670,在FPGA端硬件实现了基于肤色检测的人脸检测算法,其中包括多个IP核的设计,如CMOS图像采集模块,颜色空间转换模块,形态学滤波模块,人脸检测模块,数据传输控制模块等。最后,实验结果表明,以USB3.0为高速图像传输视频接口和以FPGA作为前端片上图像处理主控芯片的高速智能系统是可行的。论文对以FPGA为主控芯片的嵌入式图像开发和对USB3.0高速图像接口的设计,为以后高帧率智能相机的设计奠定了基础,方便了后续的智能图像系统的开发和产品化应用。
[Abstract]:High-speed intelligent image system is an embedded system device with high-speed image acquisition, intelligent image processing and high-speed data transmission. With the development of information industry, intelligent image acquisition system tends to be miniaturized, intelligent, networked and high-speed. How to make more intelligent, low-cost, multi-functional image equipment applied to people's production and life has become the focus of research day by day. Combined with FPGA embedded image development technology and ultra-high-speed data transmission technology of USB3.0 interface, this paper attempts to design a low-cost design for specific applications from the aspects of intelligence and high-speed. High-speed intelligent image system with low power consumption and better performance. In the aspect of high speed, the UVC interface for 1080p video stream transmission is designed by using the EZ-USB FX3 USB3.0 development kit of Cypress company. In the aspect of intelligence, based on the current mature theory of human face color detection, real-time detection of human face in video images on FPGA is realized. The main work of this thesis is as follows: firstly, the feasibility of face detection algorithm based on skin color is verified on MATLAB platform; Secondly, the IP core design and the USB3.0 interface design of 1080p image transmission are completed, including the design of ping-pong buffer structure to reduce the delay of high-speed data transmission. Finally, the hardware implementation of skin color algorithm model is completed. The face detection algorithm based on skin color detection is implemented by using low-speed camera OV7670, in FPGA hardware, including the design of multiple IP cores, such as CMOS image acquisition module. Color space conversion module, morphological filtering module, face detection module, data transmission control module. Finally, the experimental results show that the high-speed intelligent system with USB3.0 as the high-speed video transmission interface and FPGA as the on-chip image processing master chip is feasible. The development of embedded image based on FPGA and the design of high-speed image interface of USB3.0 have laid a foundation for the design of intelligent camera with high frame rate in the future. It is convenient for the following intelligent image system development and product application.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】