摘要
为了实现结构光对三维信息的测量,精确、快速且稳定地提取结构光光条中心是基本要求。提出了一种融合极值法和灰度重心法的结构光光条中心提取算法。首先对感兴趣的光条中心及其附近区域进行灰度增强,增加了该部分200-255灰度区间灰度的对比度。然后利用极值法找到各行列的极大值,对这些极值利用8邻域比较去噪算法,初步确定部分光条中心位置。接着利用灰度重心法计算部分光条中心位置,最后对于整条光条中心从全局角度进行多项式曲线拟合。加入噪声密度为1×10^-5噪声后处理结果与未加噪声前基本一致。该算法与形态学算法进行对比能达到亚像素级精度。
In order to realize the measurement of the three-dimensional information of the structured light, it is an essential requirement to accurately, quickly, and stably extract the center of the structured light strip. An algorithm for extracting the center of the structured light stripe is proposed, which combines the extreme value method and the gray center of gravity method. First, grayscale enhancement is performed on the center of the strip of interest and its vicinity, and the contrast of the gray scale of the portion from 200 to 255 gray scale is increased. Then use the extremum method to find the maximum value of each row and column, and use these 8 neighborhood comparison denoising algorithms to determine the center position of some strips. Then, the center position of the partial light strip is calculated by the gray center of gravity method, and finally the polynomial curve fitting is performed from the global angle for the entire center of the light strip. After adding noise with a noise density of 1×10-5, the post-processing results are basically the same as before the noise is added. Compared with the morphological algorithm, the algorithm can achieve sub-pixel precision.
作者
梅峻华
赖磊捷
Mei Junhua;Lai Leijie(College of Mechanical Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《电子测量技术》
2018年第21期8-13,共6页
Electronic Measurement Technology
基金
国家自然科学基金(51605275)项目资助
关键词
线结构光
光条中心提取
灰度增强算法
8邻域去噪
line structured light
light stripes centre extraction
grayscale enhancement algorithm
8-neighborhood de-noising
