基于机器视觉的嵌入式机械手监控系统研究与设计

发布时间：2018-02-23 05:04

本文关键词： ARM Linux 目标检测缺陷检测 Qt　出处：《浙江工业大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着塑料加工产业的快速发展，注塑机辅助设备自动化的程度越来越高。越来越多生产塑料制品的厂商都在其注塑机上装配了机械手，用于将塑料制品从模具中取出来并送到下一个加工工序上进行再加工。然而当机械手抓取异常或抓取的产品不符合要求时，就需要一个检测系统能够实时地报警并采取必要的措施，以避免重大损失。因此，开发一个机械手的检测系统具有重要的应用价值和现实意义。检测技术作为现代工农业的一个基础技术，是确保产品质量的关键。然而传统的检测技术已经不能满足现代生产的要求，而且也不能保证检测的效率与精度。如何准确有效地判断出机械手是否抓取到产品以及识别产品的优劣是目前机械手控制系统所面临的主要问题，本文正是在这种研究背景下，展开了基于机器视觉的嵌入式机械手监控系统的研究。 1、首先阐述了课题的背景和意义，其次综述了国内外的研究现状，最后介绍了本文的研究内容和论文结构。 2、在对嵌入式机械手的监控系统进行需求分析后，搭建了系统的软硬件平台。其中硬件平台采用ARM9作为主控芯片，配置了一些必要的硬件环境，如网口、串口、USB口等；在此基础上成功移植了Bootloader、kernel和filesystem，完成了软件平台的搭建。 3、针对机械手控制系统面临的两个主要问题，采用了基于机器视觉的目标检测算法和缺陷检测算法。实验证明，，该算法能够有效地判断机械手有无抓取异常并检测出有缺陷的产品，从而实现及时报警，避免了重大损失。 4、针对监控系统的要求，制定了软件的整体方案，主要包括图像采集、界面显示和数据传输三部分。其中界面部分采用Qt编写，主要实现监测信息显示和功能设置等。 5、对本文的研究成果进行了总结，并指出一些有待改善的不足之处。
[Abstract]:With the rapid development of plastic processing industry, the degree of automation of auxiliary equipment of injection molding machine is becoming higher and higher. More and more manufacturers of plastic products have assembled manipulator on their injection molding machine. Used to remove plastic products from the mold and send them to the next processing process for reprocessing. However, when the manipulator grabs abnormal or fetched products that do not meet the requirements, It is necessary for a detection system to alarm in real time and take necessary measures to avoid heavy losses. Therefore, the development of a manipulator detection system has important application value and practical significance. As a basic technology of modern industry and agriculture, detection technology is the key to ensure the quality of products. However, the traditional testing technology can no longer meet the requirements of modern production. Moreover, it can not guarantee the efficiency and precision of the detection. How to accurately and effectively judge whether the manipulator grabs the product and identifies the advantages and disadvantages of the product is the main problem that the manipulator control system faces at present. The research of embedded manipulator monitoring system based on machine vision is carried out. 1. Firstly, the background and significance of the project are expounded, then the current research situation at home and abroad is summarized. Finally, the research content and structure of this paper are introduced. 2. After analyzing the requirement of the monitoring system of embedded manipulator, the software and hardware platform of the system is built, in which the hardware platform uses ARM9 as the main control chip, and some necessary hardware environment, such as network port, serial port port, etc. On this basis, the bootloader kernel and filesystem are transplanted successfully, and the software platform is built. 3. Aiming at the two main problems of manipulator control system, the target detection algorithm and defect detection algorithm based on machine vision are adopted. The algorithm can effectively judge the grab anomaly of manipulator and detect the defective product, so that the alarm can be realized in time and the heavy loss can be avoided. 4. According to the requirements of the monitoring system, the whole scheme of the software is developed, which mainly includes three parts: image collection, interface display and data transmission, among which the interface part is written by QT, which mainly realizes the display of monitoring information and the setting of function. 5. This paper summarizes the research results and points out some shortcomings that need to be improved.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP368.1;TP241

【参考文献】