首页 | 本学科首页   官方微博 | 高级检索  
     

平行照明多目视觉检测技术
引用本文:赵阳,曲兴华. 平行照明多目视觉检测技术[J]. 红外与激光工程, 2010, 39(2)
作者姓名:赵阳  曲兴华
作者单位:天津大学,精密仪器与光电子工程学院,天津300072
摘    要:多目视觉检测技术是通过多个不同位置的视觉传感器,从不同角度获取同一目标的图像,并对所得图像进行目标匹配、差分相减等,最终清晰提取出目标信息的一种视觉测量技术.相比单目视觉方法,利用多个传感器能搜集更多的目标物信息,且在检测系统中又融入了平行光照明方式下,对称式多角度的图像采集方法,可以使不同显现力的被测物平面信息通过极高的对比度显现出来,再结合多幅图像匹配、差分规则,最终可以得到效果更佳的缺陷或轮廓信息.描述了遵照该检测方法设计的实验装置,并通过成组实验证明了其可行性.基于平行光照明的多目视觉检测方法提高了系统的抗干扰能力,使系统从根本上提高了图像处理结果的精度,相比以往软件除噪、滤波等方法更加便捷、快速、稳定,适用范围也更广.提出了利用大功率LED芯片构造平行光源的思路,该平行光源设计准直度更好,非平行杂光干扰更小.

关 键 词:多目视觉检测  平行光  差分

Multi-vision detection method based on parallel lighting
ZHAO Yang,QU Xing-hua. Multi-vision detection method based on parallel lighting[J]. Infrared and Laser Engineering, 2010, 39(2)
Authors:ZHAO Yang  QU Xing-hua
Abstract:Multi-vision detection was an image processing technology which captured images of the same object from different angles with several CCD sensors,and obtained the information of the object after image matching,segmentation and subtraction.In contrast with the method of monocular vision,the multi-vision sensors can obtained more information of the object,moreover,the detecting system used the method of captunng images from multi-angle with symmetry in the condition of parallel fighting.Finally,the defect and profile information of the object formed perfectly with the help of the method of image matching and the rules of the image subtraction.According to the technology of the image detection above,an experimental device was introduced in the paper.The groups of experiments prove that the method is quite feasible.The method of multi-vision detection based on parallel lighting greatly improves the anti-interference ability of the system,and fundamentally raises the precision of the image processing.This method is much more convenient,quick,stable and applicable to different situations compared with the traditional method of eliminating noise and filtering.
Keywords:Multi-vision detection  Parallel fighting  Subtraction
本文献已被 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号