基于机器视觉路径识别的激光加工系统的研制

发布时间：2017-12-27 01:00

本文关键词：基于机器视觉路径识别的激光加工系统的研制　出处：《南京理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：在中国制造2025计划中,中国在逐步进入"光制造"时代,而"光制造"的主要领域之一就是激光加工。在现代化的工业生产中,激光加工技术越来越普遍越来越重要,其高效性精确性以及适应多种不同加工材质的特性使得传统加工手段无法与之比拟。随着工业自动化技术的提高传统的车床加工技术开始进行升级换代,激光加工系统越来越先进,特别是与机器视觉的技术相结合,使得其精度更高效率更好智能化程度更高。传统的加工车床很少带有机器视觉,当在工作面上某位置打孔时需要在加工前手工测量计算目标点位置坐标,比较麻烦且手工测量误差较大。本论文根据所研究的激光刻蚀系统的性能指标以及功能需求来进行设计,提出了加工精度要求达到0.5mm,搭建了以STM32为核心的硬件平台,并设计了外围的扩展电路及接口。开发了上位机人机交互界面以及整个下位机控制程序,包含有文件传输、数据读取、插补算法、闭环控制、电机加减速等。同时针对上述传统加工机床的问题本论文研究了基于机器视觉的激光加工路径识别算法,使得加工系统可以自主进行目标点定位以及自主识别所要加工的路径并自动按路径进行加工,通过软件仿真和实验有效的证明了其识别精度达到了 0.1mm。最后进行了整体的调试以及测试,将各模块连接并下载程序并加工数种图案,加工过程流畅,并对加工后的图案使用游标卡尺测量以及视觉检测得到其加工的精度为0.5mm,表明可以达到所要求的性能。
[Abstract]:In the 2025 plan made in China, China is gradually entering the era of "light manufacturing", and one of the main areas of "light manufacturing" is laser processing. In modern industrial production, laser processing technology is becoming more and more common and more and more important. Its efficiency, accuracy and adaptability to various processing materials make traditional processing methods unable to match. With the improvement of industrial automation technology, the traditional lathing technology has begun upgrading. The laser processing system is more and more advanced. Especially with the technology of machine vision, it makes the accuracy more efficient and intelligent. Traditional machining lathes seldom have machine vision. When drilling at a certain location on the working face, it is difficult to manually measure and calculate the location coordinates of the target points before machining, and the manual measurement errors are relatively large. According to the performance and functional requirements of laser etching system, this paper designs the precision requirement of the laser to achieve 0.5mm. The hardware platform based on STM32 is built, and the peripheral expansion circuit and interface are designed. The upper computer's human-machine interaction interface and the whole slave computer control program are developed, including file transmission, data reading, interpolation algorithm, closed loop control, motor acceleration and deceleration, etc. At the same time, according to the traditional processing machine this paper studied laser processing path recognition algorithm based on machine vision processing system, makes the path can independently target point positioning and self identification to be processed and automatically according to the path of processing, through software simulation and experiment proved its accuracy reached 0.1mm. Finally, the overall debugging and testing, each module is connected and download the program and the number of processing patterns, the process smooth, and the processing of the processing patterns measured by vernier caliper and visual detection accuracy of 0.5mm, that can achieve the required performance.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TN249

【相似文献】