节点文献

圆形基准定位技术研究

Research on the Locating of Circular Mark

【作者】 李瑛

【导师】 刘国良;

【作者基本信息】 哈尔滨工业大学 , 控制科学与工程, 2012, 硕士

【摘要】 本文针对贴片机高速度高精度的要求,对贴片机视觉系统所用到的圆基准点定位技术进行研究。贴片机上的基准点包括机械零点、送料器两侧的基准点、PCB板基准点以及精密元件上的局部基准点。对基准点的定位主要用来补偿校正机械加工装配误差、精确确定位置偏移、获得精确的元件拾取位置和元件贴装位置。基准点的定位精度对贴片机的精确贴装起到了至关重要的作用,直接影响着贴片机的精度。根据圆定位技术的要求,对其所用到的三类算法进行了研究,它们分别是连通区域标记、改进Hough变换、亚像素边缘定位,具体描述:(1)根据定位算法对预处理技术的要求,提出了一种新型的连通区域标记算法,它结合了边缘跟踪以及线传播、区域增长的思想。实验证明,新的算法在保证标记效果的前提下降低了算法消耗时间,并且获得了顺序排列的边缘数据。另外,当存在一定强度的噪声时,仍能可靠区分不同目标。(2)针对标准圆Hough变换时间复杂度和空间复杂度都较高的缺点,提出了对标准圆Hough变换的改进。利用圆的几何特征——不共线的三点确定一个圆以及相交弦的垂直平分线交于圆心——来求取圆心,利用半径直方图求取半径。实验和理论分析证明,算法在保持了标准圆Hough变换优点的同时,提高了算法运行效率,降低了空间开销。(3)介绍了三种较为常用的亚像素边缘定位算法,并对三种算法在不同信噪比情况下的定位精度进行了实验比较,通过实验结果得到了适合圆定位的亚像素算法。通过实验证明,本文提出的圆定位算法具有较高的定位精度和较快的处理速度,达到了设计要求。

【Abstract】 In view of requirements for high speed and high precision, this paperresearched locating of the circular reference point used in vision system ofplacement machine. Reference point of the placement machine including mechanicalzero reference point, the reference point of the feeder, the reference point of thePCB board and the local reference point of some precision components. Theaccuracy of locating of the reference point, which used to correct the mechanicalmanufacturing errors, and accurately determine the location offset, accuratelydetermine component`s pick position and placement position, hasplayed the important role of the accurate placement of the placement machine, and itdirectly affects accuracy of the placement machine.Based on the technology of circle location, this paper researched threealgorithms: connected component labeling, Hough transform and sub-pixel edgelocation. The detail is as follows:(1) According to the requirements of the locating algorithm forpreprocessing technology, a new connected component labeling algorithm wasproposed. And it combines the idea of the edge track and the line spread, regionalgrowth. Experiments show that the new algorithm keeps the marked effect, andreduces the time complexity; in addition we can get the sorted edge data by thisalgorithm. The labeling of noised image was discussed.(2) For the high time complexity and space complexity of the standardcircular Hough transforms, we improve on the algorithm. The improved algorithmcalculates center coordinates of the circle using the circular geometric property, theyare circle can be determined by three non-collinear points and the perpendicularbisectors of two intersected chord intersect at the center of the circle. Then itcalculates the radius using radius histogram. Experimental and theoretical analysisshows that the improved algorithm keeps the advantages of the standard circularHough transforms, improves the algorithm’s efficiency, and reduces the algorithm’sspace complexities.(3) Three typical sub-pixel edge localization algorithms were described.And we researched location accuracy with the noise of various signal-to-noiseratios (SNR). And through the experimental result, we find the sub-pixel algorithmthat best fits the circular location.The experimental results show that the circle locating algorithm proposed inthis paper is provided with the higher precision of location and faster processingspeed, and the location algorithm of circular mark can satisfy the designrequirements.

  • 【分类号】TN405
  • 【被引频次】9
  • 【下载频次】215
节点文献中: 

本文链接的文献网络图示:

本文的引文网络