基于数字化图像处理的激光打标系统的研究

发布时间：2018-09-03 15:36

【摘要】： 随着激光器的发展,激光打标技术在激光加工的领域中应用越来越广泛,比如电子工业、汽车工业、五金、日常用品、标签技术、航空工业、仪器仪表以及广告标牌等行业。它具有传统方法(腐蚀加工、电火花加工、机械刻化加工、印刷等)难以比拟的优点,其优点在于打标效率高、精度高、质量高、热影响区小、无污染等,因此激光打标系统值得深入分析和研究。目前,激光打标的主要对象分为图像和文字,而文字可以通过其它方式转化保存为图像,因此本文主要针对以图像作为激光打标对象存在的问题进行研究。图像主要分为位图和矢量图,位图虽然内容丰富,应用广泛,但是占用空间大、打标效率低,失真大;矢量图虽然占用的空间小,失真小、效率高,但是矢量图大都依靠AutoCAD等软件绘制成,生成的图形简单,绘制时间长,较复杂的图形,要用某些软件进行轮廓勾画,过程很繁琐。针对以上的问题,本文研究了一种位图转化为矢量图的算法流程,该流程主要分为两大部分:1.图像预处理部分,图像预处理包含彩色灰度化处理、灰度变换处理、滤波处理、二值化处理;2.图像特征处理部分,图像特征处理包含图像边缘提取、边缘细化处理、边缘矢量化处理。第一部分的目的是为图像轮廓边缘提取和矢量化处理提供一个较好的初始化环境;第二部分的目的是提取完整的图像轮廓边缘,细化边缘宽度为一个像素,而后进行边缘的矢量化处理,获取图像轮廓边缘的坐标,进而能控制激光打标系统的激光器运动轨迹。鉴于目前激光打标控制的方式的优缺点,本文提出了工控机+运动控制卡控制方式构架本系统的硬件平台。运动控制卡采用了乐创自动化有限公司开发的高性能的多轴运动控制卡MPC07,其应用在系统中的主要功能分为两轴步进电机控制和激光器控制。由于该运动控制卡配备了功能强大、内容丰富的Windows驱动程序、DLL函数库,本文利用这些函数库进行系统二次开发,大大缩短了打标系统开发的时间。本论文的重点放在位图转为化矢量图的处理算法的研究,研究其流程算法中的各种处理方法的优缺点,根据不同图像特征去选择合适的图像处理算法。为了能自适应提取各种图像完整边缘,本文还针对传统边缘检测Canny算子进行一些的改进和完善。不同特征的图像使用的图像处理算法不同,因此流程中所包含的步骤的处理算法很多,其算法的适合性根据其处理的图像效果进行判断。为了验证该算法流程的可行性,本文采用面向对象的思想把位图转化为矢量图流程的各种图像处理算法集成于各个模块中,运用UML建模设计方法实现整个系统的建模,利用Visual C++6.0实现各种图像处理的算法的代码编写和整个系统集成,为激光打标系统提供友好的人机界面。
[Abstract]:With the development of laser, laser marking technology is more and more widely used in the field of laser processing, such as electronic industry, automobile industry, hardware, daily necessities, label technology, aviation industry, instruments and instruments, advertising signs and other industries. It has unparalleled advantages over traditional methods (corrosion machining, EDM, mechanical etching, printing, etc.). It has the advantages of high marking efficiency, high accuracy, high quality, small heat affected area, no pollution, etc. Therefore, laser marking system is worthy of in-depth analysis and research. At present, the main objects of laser marking can be divided into images and characters, and characters can be transformed into images by other means. Therefore, this paper mainly focuses on the problems of using images as laser marking objects. The image is mainly divided into bitmap and vector map. Although bitmap is rich in content and widely used, it occupies large space, low marking efficiency and large distortion, although vector map occupies small space, distortion is small, and efficiency is high. However, most of vector graphics are drawn by software such as AutoCAD, the generated graphics are simple, the drawing time is long, and the figure is more complex, it is very complicated to use some software to outline the outline. In order to solve the above problems, this paper studies a bitmap into vector graph algorithm flow, the flow is divided into two parts: 1. Image preprocessing, image preprocessing includes color grayscale processing, grayscale transformation processing, filtering processing, binary processing. Image feature processing includes image edge extraction, edge thinning and edge vectorization. The purpose of the first part is to provide a better initialization environment for image contour edge extraction and vectorization processing; the second part is to extract the complete image contour edge, thinning the edge width to one pixel. Then vectorization of the edge is carried out to obtain the coordinates of the edge of the image contour, and then the laser trajectory of the laser marking system can be controlled. In view of the advantages and disadvantages of laser marking control, this paper presents the hardware platform of the system based on the control mode of the motion control card of the industrial control computer. The motion control card adopts the high performance multi-axis motion control card MPC07, developed by LeChong Automation Co. Ltd. Its main functions in the system are divided into two axis stepper motor control and laser control. Because the motion control card is equipped with powerful and rich Windows driver library, this paper makes use of these function libraries to develop the system, which greatly shortens the time of development of marking system. This paper focuses on the study of bitmap to vector image processing algorithm, study the advantages and disadvantages of various processing methods in the process algorithm, and select the appropriate image processing algorithm according to different image characteristics. In order to self-adaptively extract the complete edge of various images, this paper also improves and perfects the traditional edge detection Canny operator. Different feature images use different image processing algorithms, so there are many steps involved in the process. The suitability of the algorithm is judged according to the image effect. In order to verify the feasibility of the algorithm flow, this paper integrates various image processing algorithms of bitmap into vector chart flow, and uses UML modeling design method to realize the modeling of the whole system. Visual C 6.0 is used to realize the coding of various image processing algorithms and the integration of the whole system, which provides a friendly man-machine interface for the laser marking system.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：TP391.41

【参考文献】