摘要
针对复杂的工业现场环境以及打印机打印质量水平的限制,对于工业现场的摄像机标定,采用直接提取平面模板角点的方法往往出现失败.现对不能顺利提取角点的标定板图像采用一种基于区域分割的图像处理方法,首先采用均值滤波的方法消除打印机打印质量问题,然后对图像像素按照一定的准则采用区域生长算法得到区域边界,最后与滤波图像进行相减得到所需图像.在一定程度上,这种方法能够消除工业现场环境的干扰,达到精确提取标定板角点的目的.通过分析摄像机模型和几种坐标系之间的转换关系,利用标定板完成了摄像机标定.结果表明,采用该方法,能够准确、方便地获得工业现场中摄像机的内外参数.
Because of the complicated environment of the industrial site and the limited quality of the printer printing,the failure often occured in extracting the model plane's corner directly for the camera calibration in the industrial site.Now,an image processing method based on the regional segmentation was used for the calibration plate image whose corner was not extracted smoothly.Firstly,we used the average filter to eliminate the quality problems of the printer printing,and then region growing algorithm was used in accordance with certain rules for the image pixels,obtaining the regional boundary,finally,the filtered image subtracts the boundary and then the image Obtained.To a certain extent,the method eliminates the interference of the environment of the industrial site,achieving the purpose of extracting the corner of the model plane precisely.We analyzed the camera model and several conversions between coordinate systems,and then use the model plane to finish the camera calibration.The result shows that the internal and external parameters of the camera can obtained in the industrial site accurately and conveniently.
出处
《中国计量学院学报》
2011年第3期246-250,290,共6页
Journal of China Jiliang University
