节点文献

激光打标中图像处理的研究

Research on Image Processing of Laser-Marking

【作者】 刘晓敏

【导师】 林树忠;

【作者基本信息】 河北工业大学 , 机械制造及其自动化, 2005, 硕士

【摘要】 论文基于YAG激光打标机控制器原理及图像处理理论,开发研究针对BMP图像文件格式适用于扫描式打标的激光打标软件。该软件的主要任务是把目前比较成熟的图像处理方法应用到激光打标机中,得到良好的图像效果,为系统后续打标操作提供图像数据。论文主要偏重于图像处理开发自己的软件,软件主要包括两大部分,即图像处理部分和执行打标部分。 一个BMP位图文件通常包括位图文件头、位图信息头、调色板、位图数据四个部分,基于Visual C++面向对象编程方法,构建设备无关位图CDIB类库,完成DIB文件的读写操作及图像显示和调色板操作,提供位图宽度、高度、颜色数目等相关信息。 根据打标对图像的要求,对彩色图像进行灰度化处理,通过对图像进行灰度变换,改变原始图像所占据的灰度范围达到了良好的视觉效果。几何变换改变了图像中各物体之间的空间关系,组合图像的平移、镜像、转置、反转、旋转等处理,可以实现各种各样的图像几何变形,达到对物体几何位置及不同几何形状的要求。 对于激光打标系统,其控制的激光状态只有两种,即出光与不出光,对应图像就是黑与白,因此就要求图像通过阈值处理来得到二值图像,根据不同的图像选用合适的阈值实现灰度图像的二值化处理。 边缘检测首先检出图像局部特性的不连续性,然后再将不连续的边缘像素连成完备的边界,得到连续图像,针对不同的图像采用不同的边缘检测算子提取图像特征。 图像在形成、传输、接收和处理的过程中,会产生噪声,对图像进行去除噪声处理,可以减少图像的冗余信息。通过腐蚀膨胀可以较好的去除二值图像的噪声。而对图像进行细化突出了形状特点并减少了冗余信息量。 基于PC、D/A转换器和双振镜扫描系统搭建激光打标机控制系统,打标时,通过D/A转换器将图像坐标数据转换为模拟电压值,控制振镜的转动确定激光光路,并设定了延时。 最后,论文介绍了本系统的硬件实现和软件算法,给出了图像软件程序调试及打标的结果。

【Abstract】 According to the principle of the controller of YAG marking machine and the principle of image processing, the laser marking software , which is suitable to BMP image for scanning marking, is built in this paper . The main task of this software is to receive the good effect of the image and to obtain the datum to following marking process through using the successful methods of image processing during the laser marking. The paper particularly stresses on the software which includes image processing and the distilling datum to laser marking.Usually, a BMP file includes four parts, bitmapfileheader ,bitmapinfoheader, RGBQUAD and the data of image.The CDEB class that is defined in this paper is based on the object oriented programming method of Visual C++ in order to accomplish reading, writing and display of the image files ,and offers width, height and color of the files.The color image is changed to gray image in order to the requisition of the laser marking. Gray range of the original image is changed by gray transformation in order to get the well vision effect. The spatial relationship between the objects of the image will be changed by geometrical transtormation. The geometrical distortion of the image will be received by the combination of move transform, mirror transform, transposition, reversal transform, rotation transform and so on.There are two kinds of laser states in laser marking system, laser-open and laser-close that correspond to black and white in one image. So binary image is required by binarization of image with appropriate threshold of vary original image.The discontinuity of image part feature is firstly detected in the processing of edge detection, then, consecutive image is obtained with pretty edge detecting operator to link the image edge.The noise will be produced in the process of image forming, image transmitting, image receiving, the redundant image information can be reduced in the process of smoothing. It is proved that dilition and erosion are better methods in getting rid of the noise of binary image. The figure characteristic will be stressed and the redundant information will be reduced in the process of thinning.Based PC, D/A convertor and 2-dimension galvanometer scanning system, laser marking control system is built. Image coordinate data are transformed into simulant voltage by D/A convertor. At the same time D/A convertor control the galvanometer to deflexion. Delay is considered in the control process.Finally, the paper introduces the configuation of the whole hardware and software in this systerm. The result of the image software debugging and the result of marking are especially displayed.

  • 【分类号】TG665
  • 【被引频次】19
  • 【下载频次】884
节点文献中: 

本文链接的文献网络图示:

本文的引文网络