摘要
为了提高自准直仪的测量精度和降低人眼瞄准的劳动强度和主观误差,提出一种基于机器视觉的二维自跟踪视觉准直测量系统。该系统由光学准直仪、CCD和图像采集卡等组成。以面阵CCD代替人眼对自准直仪的十字分划板图像进行瞄准;采用图像不变矩自动搜索算法和hough变换、亚像素细分等方法实现对小角度的自动精确测量。实验结果表明,该系统测量范围达到±15',最小分辨力优于0.1″,与MJS双频激光测角系统比对误差在0.6″以内,测量精度高于MJS,而且操作方便,易于推广。
In order to increase the measurement precision of an autocollimator and decrease the fatigue of eye during aiming target and visual error, a two-dimension automatic tracking vision autocollimator based on machine vision is presented. The system is composed of an optical autocollimator, a CCD and image collector board etc. The system carries out automatic and precision measurement of small angles by using CCD rather than eye to aim cross reticule image on the autocollimator, and by means of auto-search algorithm with image moment invariants, Hough transformation, sub-pixel subdivision, etc. Experiments demonstrate that this system has large measurement range up to ±15' with high resolution of 0.1' The error in contrast to the MJS double-frequency laser goniometric system is less than 0.6'. This system has higher precision than MJS with merits of convenient operation and easy popularization.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第5期55-57,68,共4页
Opto-Electronic Engineering
基金
哈尔滨工业大学校基金资助项目(HIT.MD.2001.18)
关键词
自准直仪
自动跟踪
图像不变矩
机器视觉
模式识别
Autocollimators
Automatic tracking
Image moment invariants
Machine vision
Patten recognition