高斯混合分布激光中心线提取方法  

Extraction method for laser stripe centerline with Gaussian mixing distribution

作　　者：刘巍[1] 张驰[1] 刘阳[1] 王灵丽[1] 樊超楠[1] 贾振元[1] 

机构地区：[1]大连理工大学机械工程学院,辽宁大连116024

出　　处：《激光与红外》2015年第11期1397-1402,共6页Laser & Infrared

基　　金：国家自然科学基金项目(No.51227004)资助

摘　　要：针对激光辅助立体视觉测量中的非高斯非对称分布激光条的匹配中心线提取精度较低的问题,提出一种基于高斯混合模型的激光中心线提取方法。首先分析了激光散斑对图像的影响,选择均值滤波方法有效去除激光散斑噪声;然后利用最大类间方差(OTSU)阈值分割方法对光条位置进行粗定位;最后,利用本文提出的高斯混合模型提取激光条亚像素中心线,该模型可准确地描述激光条横截面光强分布特性,从而能够实现光条中心极值点的高精度提取。实验结果表明该方法能够高效稳定的提取激光条中心线。To solve the problem of low extraction precision of laser stripe centerline distributed in a non-Gaussian and non-uniform manner in stereo vision measurement,an extraction method of laser stripe centerline based on a Gaussian mixture model is proposed. First of all,the influence of laser speckle on the image is analyzed,and then laser speckle noise is effectively removed by mean filtering. Then position range of the laser stripes is located by threshold segmentation based on the OTSU method. Finally,the sub pixel centerline of laser stripes is extracted by Gaussian mixture distribution model. A Gaussian mixture distribution model can accurately describe the intensity distribution of laser stripes cross section,thus can achieve a high-precision extraction of extreme points at the center of the laser stripes. Experimental results show that this method can efficiently extract the centerline of the laser stripes.

关 键 词：立体视觉测量 图像处理 中心提取算法 激光光强分布 高斯混合模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]