基于虚拟仪器的智能视觉识别机械手系统

A Virtual Instrument based Manipulator System with Intelligent Visual Identification

目的:设计与搭建适应检测物体更换较快应用场合的可自行学习并更换对象模板的视觉智能机械手系统。方法:利用LabVIEW开发平台和NI Vision软件包,选择合适的硬件设备搭建视觉智能机械手系统:采用调用DLL的方式配置相机参数及控制其进行图像采集;并对采集到的图像进行LUT,阈值分割和形态学变换,消除一定光照干扰等环境变化影响,使其适于下一步分析;通过Geometric匹配及对特征区域的灰度值判断,实现物件的识别与检测。同时针对物件识别需要的信息设计与完成相应的模版学习与更换程序。结果:实现了对目标物体的视觉识别及控制机器手的抓取,视觉识别准确率达到97%以上,其特色还在于可以在应用实践中实时实地方便地进行自主学习与制作对象模板,在目标频繁变更的情况下快速实现对不同物体的视觉识别与抓取控制。结论:采用LabVIEW开发平台和NI Vision软件包可搭建出具有视觉智能识别功能,且识别准确率较高,可便捷地学习与更换识别对象的智能机械手系统,具有应用价值。

Objective：To build a visual sorting system with self-lealnable ＆ replaceable template and use it for template fast changing applications.Methods：The industrial camera was configured and controlled by calling DLL. The LUT, threshold seg mentation, and morphological transformation were performed to eliminate the environmental interferences and make the image suitable for analysis. Geometric matching and Grey value detection in characteristic area were combined for visual sorting, and a template built program was also writen. All these things were based on LabVIEW platform and NI Vision software package.Re suits： The system can perform online detecting and sorting and has accuracy better than 97%. It also has templates of quick self-learning and changing.Conclusions： By using IabVIEW and NI Vision, a visual sorting system with high accuracy and tem plates of quick learning and changing has been built. It＇s enough for the application.