摘要：论文主要研究了自动物料分拣机器人中的视觉定位技术。自动物料分拣机器人对于目标物料的识别是自动分拣视觉系统的重点与难点，本文主要通过研究机器人视觉图像的预处理，边缘检测以及Hough变换来解决机器人视觉中的目标识别与定位问题。31525
研究过程首先是标定了相机，即将世界坐标系与图像坐标系通过直接线性转换连结在一起，这是进行视觉定位的基础。然后采集原始图像，采用CCD摄像机拍摄了一个轴承的原始图像，并输送到计算机当中进行图像的预处理。预处理采用Matlab软件，处理的内容包括了图像灰度的变换，直方图的均衡化，图像噪声的滤波以及二值化等，比较了不同的算法，并取其中较好的中值滤波来处理原始图像。接着，进行了目标边缘的检测，比较了基于一阶和二阶以及Canny算法等共6种算法的处理效果，选取了处理效果最好的Canny算法来进行图像目标的边缘检测。提取了目标边缘后，便进行了基于Hough变换的目标识别与定位，即将图像空间中的圆特征转为参数空间中的参数特征，进行累加投票计算，计算出参数空间中得到的最大参数，也就是得到了图像空间中圆的圆心和半径，从而实现了目标的识别与定位。
毕业论文关键词： 机器人视觉；自动分拣；图像预处理；边缘检测；Hough变换
Research on visual positioning technology of automatic material sorting robot
Abstract: This paper mainly studies the visual positioning technology in the automatic material sorting robot..For the Automatic material sorting robot, the identification of target material is the key points and difficulties of the visual system of automatic sorting, in this paper, through the research on robot visual image pre processing, edge detection and Hough transform to solve robot vision target recognition and location.
    The research process is calibrated camera firstly, the world coordinate system and the image coordinate system through the direct linear conversion link together, which is the basis for visual positioning. Then the original image was collected and the original image of a bearing was captured by CCD camera, and the image preprocessing was carried to the computer.. Pretreatment with MATLAB software, content processing includes image gray transform, histogram equalization, image noise filtering and binarization, different algorithms are compared and the better of median filter to process the original image. Next, edge detection, comparing the processing results based on the first-order and second-order and Canny algorithm, a total of six kinds of algorithm selects the best treatment effect of Canny algorithm for target image edge detection. After extracting the edges of target, the target identification and orientation based on Hough transform is circle feature in image space in turn for parameter space of characteristic parameters of cumulative voting calculation, calculate the maximum parameters obtained in the parameter space, also is the center and the radius of the round in the image space, in order to achieve the target recognition and location.
Keywords: Robot vision; automatic sorting; image preprocessing; edge detection; Hough transform
目 录
摘要 Ⅰ
Abstract ..Ⅱ
第一章 绪论 ...1
  1.1  研究背景..1
  1.2  相关研究状况..1
  1.3  分拣系统研究成果..3
  1.4  论文的主要工作与论文结构..5
第二章 相机标定.6
  2.1  相机的针孔模型..6
  2.2  相机标定方法..8
2.2.1  直接线性法...9
2.2.2  Tsai法..11
2.2.3  张正友法...11
  2.3  本章小结..11
第三章 图像采集及预处理.12
  3.1  图像采集..12
  3.2  图像预处理..13 自动物料分拣机器人的视觉定位技术研究:http://www.751com.cn/zidonghua/lunwen_27721.html