【作者】 李丽宏；

【导师】 孙立新；

【作者基本信息】 河北工业大学 ， 机械制造及其自动化， 2003， 硕士

【摘要】 随着科学技术的发展，许多领域越来越迫切地需要微型系统或微动系统，如生物细胞、聚合物的各种操作、微外科手术、扫描探针显微镜SPM、光纤对接和微细加工等；而且随着微技术的不断发展，以形状尺寸微小、操作尺度极小为特征的微机械已成为人们从微观角度认识和改造客观世界的一种高新技术；微机械技术还有望成为研究纳米技术的重要手段，因此微机械被列为本世纪末10大关键技术之首，并受到各工业发达国家的高度重视。由于微操作工具和被操作对像的几何尺度及操作空间的限制，一般操作成功率低，制约了这项技术的普及和应用，为此本文提出了双目显微立体视觉技术的研究，借助于机器视觉来提高微操作的效率和精度。本文的主要工作和创新点包括： 对显微立体视觉系统的光路进行了研究，并将单CCD显微系统改装成双CCD显微立体视觉系统，并对此系统的数学模型进行了研究。 对图像的预处理进行了全面地研究，如图像的灰度化、平滑化、阈值选择、边缘检测、腐蚀、膨胀、轮廓提取、轮廓跟踪、图像分割、不变矩、Hough变换等，目的是产生一幅计算机易于识别和理解的图像，同时为立体匹配奠定了基础。本文采用了自动阈值选择技术，使阈值的选择更合理；对不变矩进行了优化，首次提出了旋转投影不变矩方法，适合粘连字符地识别。 灰度匹配是利用图像的灰度对两幅图像进行立体匹配，以相关函数为判断依据对图像进行立体匹配。本文利用零交叉点作为特征点对图像进行特征立体匹配，这种匹配方法对对比度和光照的变化不敏感，同时在匹配过程中也运用了多个约束条件：如外极线约束、连续性约束等，从而提高了图像的匹配速度。 对该系统进行了大量的实验，证明了该系统是合理的；对图像进行预处理，产生了一幅计算机易于识别和理解的图像，同时使计算量减低，计算速度提高，满足了图像处理的实时性；对图像的灰度匹配进行了实验，结果表明，本文采用的灰度匹配方法计算结果准确；本文的特征匹配方法具有稳定性，同时在匹配过程中使用了约束条件，从而匹配速度很快。总之，本文对图像预处理和立体匹配都进行了深入研究，并取得了一定的成果。更多还原

【Abstract】 With the development of science and technology, the need of micro system is more and more urgent in many technical fields, such as various operation of cell and polymerized substances, micro surgery, scanning probe microscope(SPM), butting optical fiber, fine manufacturing etc. With the development of microtechnology, micro machine which has the character of micro size or micro motion is new high technology from microcosmic point of view understanding and reconstructing the world. Micro machine technology is important means for researching nanotechnology, so micro machine technology is the first of ten key technology at the end of century and is regarded importantly by developed countries. The ratio of successful operating is often low due to small size of micro tool, operated objects and working room. It is very difficult that the micro operation technology becomes widespread under the circumstances. In order to improve micro operation efficiency and precision with machine vision, this paper is researching dual CCD micro stereo vision technology. The main work and innovative ideas includes:The paper researches light route of the micro stereo vision system. The single CCD vision system is changed to the two CCD micro stereo vision system. The paper researches the math model of the dual CCD micro stereo vision system.The image pretreatment is researched widely which includes graying, smoothing, the choice of threshold value, edge detecting, erosion, dilation, outline detecting, outline tracing, image segment, moment invariants, hough transform etc. The aim of the image pretreatment obtains the image which computers easily identify and understand and establishes the base of stereo match. This paper adopts auto-threshold value technology which can select better threshold value. This paper puts forward rotating projection moment invariants which optimizes moment invariants and identifies conglutinate letters.The paper processes gray match with the image gray and on the basis of correlative function. The paper processes feature stereo match with zero crossing. This stereo match method is not sensitive to the change of contrast and illumination. At the same time, the paper applies restriction such as epipolar restriction and sequential restriction which can improve the velocity of the stereo match.The paper does many tests which test the system is reasonable. The image pretreatmentnot only offers the image which computers easily identify and understand but also count quantity decreases and the velocity of count improves which contents the image realtime processing. The gray match is tested which tests the method is accurate. The feature match in this paper is stable and applies restriction which improves the velocity of the match. In a word, this paper researches the image pretreatment and stereo match deeply which acquire some fruit.更多还原