摘要
为了提高潜孔钻机定位的精度,提出采用双目立体视觉技术获得潜孔钻机定位过程中孔位标识的三维位置信息,用于指导潜孔钻机进行自动钻孔定位.首先利用双目摄像机获取含有孔位标识的图像对,然后采用隶属度函数为梯形函数的模糊阈值分割方法对孔位标识进行识别分割,利用图像的不变矩原理求取孔位标识的质心,并通过三角测量原理求取孔位的空间位置信息.最后,在潜孔钻机试验台上进行了基于双目立体视觉系统的自动定位试验.结果表明,所采用以灰度直方图为基础的模糊阈值分割能够实现对孔位标识有效信息的提取,且定位误差小于25 mm.
In order to improve accuracy of down-the-hole (DTH) drill position, binocular stereo vision was adopted to obtain a 3D location measurement of holes marker for DTH position. A pair of stereo ima- ges with holes marker were obtained by stereo cameras to be segmented by fuzzy thresholding method based on trapezoidal membership functions. The holes marker centroid was extracted by image invariant moment principle. The holes marker 3 D locations were obtained by triangulation principle. On DTH test- bed, the autonomous position experiments were implemented based on binocular stereo vision system. The experiment results show that the holes marker can be efficiently segmented by fuzzy thresholding method based on trapezoidal membership functions with position error less than 25 mm.
出处
《江苏大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2015年第1期65-69,共5页
Journal of Jiangsu University:Natural Science Edition
基金
国家自然科学基金资助项目(51175518)
关键词
潜孔钻机
双目立体视觉
模糊阈值分割
质心提取
三维位置信息
DTH drill
binocular stereo vision
fuzzy threshold segmentation
centroid extraction
3D locations