节点文献

多传感器融合智能检测机器人的研究及应用

Research on and Application of Multi-sensors Fusion Intelligence Inspector

【作者】 刘焕军

【导师】 王耀南;

【作者基本信息】 湖南大学 , 控制理论与控制工程, 2004, 硕士

【摘要】 视觉是人类最完善的感知系统之一,而模拟视觉的机器视觉也是人工智能领域中最活跃的分支之一。多传感器融合智能检测机器人是以机器视觉技术为基础,光机电一体化的智能检测设备。目前,我国机器视觉技术的研究和应用仍处于初级阶段,与发达国家相比还有不小的差距。对多传感器融合智能检测机器人的研究不仅符合市场对此类大型智能检测设备的需求,也有利于促进机器视觉技术的研究和应用。 本文在研究国内外相关资料的基础上,提出了一种多传感器融合智能检测机器人的总体设计方案,并详细介绍了各部分的结构和功能。该多传感器智能检测机器人的电气控制系统采用了分层控制体系,主控计算机负责管理和通讯,而PLC负责底层控制。同时为了能在高速流水生产线上平稳分离物品,还研制了一种新型击出器。该击出器采用了阶梯式击打的方法,即使击出对象在高速运动中也不易翻倒。 本文重点论述了多传感器融合智能检测机器人中的机器视觉系统的设计。具体讨论了摄像机和图像采集卡的选择原则,并确定了智能检测机器人的摄像机和图像采集卡。光源及照明方案对机器视觉系统具有重要意义。本文根据实际的需求设计了系统的照明方案,保证了机器视觉系统获取图像的质量。本文还讨论了如何进行机器视觉系统的软件开发的问题。 多传感器融合智能检测机器人具有瓶身检测、瓶口检测、瓶底检测和液位检测等检测功能。为了实现这些功能,针对实际检测的特点,在分析检测对象的基础上,提出了一套实用的检测算法。经过运行表明,在多传感器智能检测机器人上使用这些检测算法,检测速度可达10瓶/秒,平均检测的准确率超过95%。 根据本文研究的各项内容,已经研制成了试验样机,并在样机上对本文所述的理论和算法进行了大量的实验,证明了设计和研究结果的有效性,为今后的进一步研究和开发打下了良好的基础。

【Abstract】 Vision is one of the most perfect sensation systems of human, and the machine vision, which can simulate human’s vision, is also one of the most active branches in the domain of artificial intelligence. The multi-sensor fusion intelligent inspector based on machine vision technology is the intelligent inspection equipment integrated of optical, mechanical and electronic functions. At present, the research and application of machine vision technology is still in its initial stage in China; compared with the developed country it also has much disparity. Studying the intelligent inspector based on multi-sensor fusion not only conforms to the market demand of this kind of large-type intelligence inspector, but also is advantageous to the promotion of the research and application of the machine vision technology.Based on the studies of the related data from at home and abroad, this article presents the system design plan for the intelligent inspector based on multi-sensor fusion, and in detail introduces the structures and the function of each part. Multi-layer control system is used in the electric control system of the intelligent inspector based on multi-sensor fusion. The host computer is responsible for higher layer management and communication, while PLC is responsible for lower layer control. In order to separate goods from the high-speed production line steadily, one kind of new ejector is developed. The ejector adopts the method that hits goods with a ladder-type stroke. Thus, the ejected object is not easy to turn over even if it is on the high- speed movement.This article emphasizes on elaborating the design of the machine vision system in the intelligent inspector based on multi-sensor fusion. Specifically, it discusses the selection principle of the camera and the image capture card, and determines the use of intelligence inspector’s cameras and image capture card. Since the light source and the illumination plan is of a vital significance to the machine vision system, this article designs the system illumination plan according to the actual demands, thus ensuring the image quality gained by the machine vision system. This text has also discussed how to carry on the software development of the machine vision system.The intelligent inspector based on multi-sensor fusion has the examination functions for bottle body, the bottle mouth, the bottle bottom and fluid position. To achieve these functions, with the consideration of the actual inspection features, this article puts forward to a set of practical examination algorithms on the basis of analyzing inspected targets. The operation indicates that, by using these examination algorithms in the intelligent inspector based on multi-sensor fusion, the examination speed can reach 10 bottles/second, and the average examination accurate rate surpasses 95%.According to studies in this article, the experimental equipment has already been developed. The theory and the algorithms stated in this article have been put into use on the massive experiments on the inspector. The results of the experiments have proven the validity of the design and the findings. So, it has laid a good foundation for further research and development in the future.

  • 【网络出版投稿人】 湖南大学
  • 【网络出版年期】2004年 04期
  • 【分类号】TP242
  • 【被引频次】3
  • 【下载频次】562
节点文献中: 

本文链接的文献网络图示:

本文的引文网络