嵌入式千兆网测量相机的设计与实现
发布时间:2018-03-18 12:17
本文选题:运动捕捉 切入点:spartan6 出处:《大连理工大学》2015年硕士论文 论文类型:学位论文
【摘要】:现代动漫制作技术中的光学式运动跟踪系统,前端相机负责采集运动图像,后端计算机负责根据图像实时对标记点进行识别、跟踪和空间坐标的计算,需要传输和计算的数据量非常大。尤其对于高分辨率高帧多目标图像的情况,更是对图像数据传输和计算结果实时性的巨大考验。为了简化光学式运动跟踪系统的结构,结合图像测量算法,本文设计并实现了一种嵌入式千兆网测量相机。该系统通过图像测量算法,在测量前端进行第一次坐标计算,然后通过千兆网将运算结果传输到计算机上,最后进行三维合成,简化了系统结构,提高了系统性能。系统以FPGA为核心控制器,选择了Xilinx公司的Spartan6芯片;利用嵌入式软核MicroBlaze设计了片上系统,用于实现图像测量算法;利用千兆网作为传输链路,选择三态以太网MAC核(TEMAC)和88E1111高速以太网收发器实现千兆网传输,将运动物体的空间坐标结果发送到上位机;利用CMOS图像传感器实现图像采集,选择了Micron公司的MT9P031,图像分辨率为640×480,帧率最高可达123帧/秒。图像测量算法采用一种基于直线段的连通域快速检测算法,首先在FPGA硬件电路内进行滤波、二值化和提取边缘坐标信息,随后片上系统根据边缘坐标信息实时计算数字图像的多个标志点的空间坐标,最后将空间坐标发送到上位机。上位机的程序采用C++语言基于MFC进行开发,利用WinPcap抓取网络数据包,实现上位机和下位机之间的通信,上位机作为客户端,下位机作为服务器。上位机可以实现对下位机的参数设置、命令控制,实时图像显示和坐标轨迹显示。测试结果表明,系统完成了对图像的采集和传输,实现了对运动目标坐标的计算。该系统运行稳定,测量数据精度较高,能够实现多运动目标的实时跟踪,具有很好的应用价值。
[Abstract]:The optical motion tracking system in the modern animation production technology, the front-end camera is responsible for collecting the moving image, the back-end computer is responsible for the real-time identification, tracking and the calculation of the space coordinates according to the image. The amount of data that needs to be transmitted and calculated is very large. Especially for high-resolution and high-frame multi-target images, it is also a great test for image data transmission and real-time computing results. In order to simplify the structure of optical motion tracking system, Combined with image measurement algorithm, this paper designs and implements an embedded gigabit network measuring camera, which carries out the first coordinate calculation in the front end of the measurement by image measurement algorithm, and then transmits the result of the operation to the computer through the gigabit network. Finally, 3D synthesis is carried out, which simplifies the system structure and improves the system performance. The system takes FPGA as the core controller, chooses the Spartan6 chip of Xilinx Company, and designs a system on chip using embedded soft core MicroBlaze to realize the image measurement algorithm. Using the gigabit network as the transmission link, the three-state Ethernet MAC core (TEMAC) and the 88E1111 high-speed Ethernet transceiver are selected to realize the transmission of the gigabit network, the results of the spatial coordinates of the moving objects are transmitted to the upper computer, and the image acquisition is realized by using the CMOS image sensor. The MT9P031 of Micron Company is selected. The resolution of the image is 640 脳 480, and the frame rate can reach 123 frames / sec. The image measurement algorithm adopts a fast detection algorithm based on the straight line segment in the connected domain. Firstly, it is filtered in the FPGA hardware circuit. Binarization and extraction of edge coordinate information, and then the system calculates the spatial coordinates of multiple mark points of digital image in real time according to edge coordinate information. Finally, the space coordinate is sent to the upper computer. The program of the upper computer is developed with C language based on MFC, and the communication between the upper computer and the lower computer is realized by using WinPcap to capture the network data packet, and the upper computer is used as the client. The lower computer serves as the server. The upper computer can realize the parameter setting, command control, real-time image display and coordinate track display of the lower computer. The test results show that the system has completed the acquisition and transmission of the image. The system has the advantages of stable operation, high accuracy of measurement data, real-time tracking of multi-moving targets, and good application value.
【学位授予单位】:大连理工大学
【学位级别】:硕士
【学位授予年份】:2015
【分类号】:TP391.41
【参考文献】
相关期刊论文 前1条
1 孟华;谢小飞;丛培超;;多运动目标高速实时跟踪算法的实现[J];电视技术;2012年03期
相关博士学位论文 前1条
1 刘瑞;运动捕捉数据智能处理算法研究及应用[D];大连理工大学;2012年
,本文编号:1629538
本文链接:https://www.wllwen.com/wenyilunwen/dongmansheji/1629538.html