摘要
针对工业生产中流水线高速移动的环境条件下,工作台上同时存在多个运动的零件等工况,提出一种综合的机器视觉快速识别与定位零件的方法。创新性地通过结合不同图像特征提取方法性能差异,形成级联分析算法,提高了算法效率,也保证了零件被准确识别和定位的准确性,以满足工业生产中的实际需要。系统可以对新零件进行训练,适应不同应用场景的需求,并在模拟的实验平台上进行了大量的测试,显示其在实际工业环境中应用的前景。
Under the environment and condition of high-speed moving assembly line in industrial production,there has the working condition that several different types of moving parts are on the worktables together.To address this,we put forward a comprehensive method of fast recognising and locating the parts with machine vision.The method innovatively combines the performances difference of different image feature extraction methods and forms cascades analysis algorithm,this improves the algorithm efficiency,and ensures the accuracy of the recognition and localisation of the parts,so as to meets the actual demands of industrial production.The system can train the new parts in order to adapt for the needs in different application scenes.The method has been tested a great deal on simulative experiment platform,its prospects in practical industrial environment is exposed.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第11期153-155,共3页
Computer Applications and Software
基金
上海市科委重点科技攻关项目(08511501303)
关键词
高速移动背景
零件检测
图像特征
级联处理
Rapid moving background Part detection Image feature Cascade processing