【作者】 陈洪涛；

【导师】 程光明；

【作者基本信息】 吉林大学 ， 机械电子工程， 2007， 硕士

【摘要】 测量物体的运动参数，常规的方法主要是采用各种位移、速度、加速度传感器。但如何根据动念图像的分析，来确定物体的运动参数，目前得到了广泛的关注。本文的目标是利用机器视觉的方法来实现对运动机构的运动参数的测定，从而达到非接触式测量的目的。使用CCD摄像机采集进来运动机构的图像序列，在程序中设置了定时器，利用模板匹配技术对运动机构中设定的目标实行实时性的在线识别和跟踪，从而得到它在CCD采集进来的图像序列中每次采样时的坐标值。利用目标点的相临坐标值，采用数学的算法从而在线得到各种运动参数的曲线。针对实时性的要求和机构运动的连续性特点，模板匹配算法采用了全局匹配与小区域匹配相结合的策略。这样就使得运算量大大减小，加快了识别速度，实现了实时性的跟踪识别。通过实验可测得本文试验装置全局匹配一次的时间为700ms，小区域局部匹配一次的时间为20ms。从而可以设置定时器的采样时间间隔为30ms，能达到运动参数在线实时测试的要求。更多还原

【Abstract】 The conventional movement survey mainly is uses each kind of displacement, the speed, the acceleration instrument carries on surveys these parameter of movement. but How acts according to the dynamic image or the time-variable image sequence analysis to determination three-dimensional space mobile structure or the movement between object and viewer’s parameter. These questions are in the current machine vision research area extremely popular topic. The purpose of this paper is using the method of machine vision to measure the motion parameter of motion framework, consequently get the purpose of non-touching measure.In recent years unceasingly along developed with the machine vision and the imagery processing technology, The movement goal track and the recognition also obtained widespread application and Dynamic image processing and the analysis receives day by day takes. Specially under movement image background, The movement goal recognition and the track becomes important research topic because of its widespread application prospect. But the template match technology also obtained widespread application in the target identification track domain. It uses the known image to take template to ask the image and template similarity through Moving template in the unknown image.When finding the similarity biggest spot it is to match success. Thus can obtain the goal which the template represents in the big image whether has or existence position. Under such technical background, Pasting in the motion symbolizes and Choosing a representative spot to take we track target point on this sign, Gather the card gathering motion with the CCD camera and the image dynamic image sequence. Through the goal unceasing recognition and tracks,we can obtain this goal in each sampling time coordinate figure. using the mathematics algorithm, Then we can obtainthis goal path and the speed, parameter of movement and so on acceleration curve. This test method may realize the non-contact type to survey and equipment is simple. This article in the above conception, has carried on a series of work and the exploration.1:The system composition of this article experimental: This article experimental hardware system is mainly composed of photographs by CCD and the image gathering card . The function of the CCD camera is obtaining the job site simulation image and the function of the image gathering card is the simulation service pattern video signal which photographs the CCD camera carries on the quantification and the sampling, transforms for digital image,then deliver to the memory in the computer to supply to process. In this experimental system, the crank rocker organization is used to take experimental motion. In order to be advantageous to the goal recognition and the track, has glued a circle marker on the rocking bar, this sign center of circle took we distinguish with track goal.2: This algorithm and the software plan of the article: we can obtain this sign central point’s coordinate figure in each sampling time By carrying on the dynamic recognition unceasingly using the template match technology to this organization surface circle marker and tracks.According to the neighboring two coordinate figures, we can finally obtain this target point path curve, the speed and the acceleration curve and complete the dynamic image test duty. The sampling time-gap hypothesis is established through the timer in the procedure to complete. The timer sampling time may change as necessary in the procedure according to the situation. Theoretically speaking, this time hypothesis smaller , obtaining each kinematics parameter curve more precise. Finally we can achieves online real-time to dynamic image test.The article Proposed the overall situation match and the small region match unifies strategy according According to the crank rocker organization movement continuity, also the path existence rule, characteristic which can be forecasted. When namely in carries on the target identification and the track which matches based on the template, does not carry on the overall situation the search match to distinguish, but is only when the system first goal match recognition, carries on the overall situation match to distinguish the goal initial point. Later each time will match will distinguish, above the recognition position takes the initial point. In carries on take this initial point as in a central certain distance scope small region matches and distinguishes. Like this we can caused the operand to reduce greatly, sped up the recognition speed, has realized the timely track recognition. 3: Conclusion:This system gathers overall situation whose picture size is 400x300 picture element,and when partial match the small region size is 80x60 picture element. It observed that the overall situation match time is 700ms and the small region match time is 20ms with the test installation through the experiment.. This kind of test method realize to test the parameters of movement online timely test very sucessfully. There is a good precision and the method is extremely simple, The application prospect is very wide.更多还原