摘要
在基于线结构光的视觉测量系统中,激光光条中心的提取精度是影响系统最后精度的关键因素。目前常用的光条中心提取方法都没有去除光条上的干扰噪声,在分析现有方法的基础上提出了一种利用B样条迭代去除光条噪声的方法。以光条上的点作为控制顶点拟合B样条曲线,在同一位置下用B样条曲线上的点取代噪声点,这样反复修改噪声点的位置使其逐渐逼近实际曲线,从而达到去除噪声的目的。利用重心法提取去噪后光条的中心,着重分析了参与运算的像素的取值范围。通过对比实验证明该方法的提取精度远高于其它方法,从而验证了去除噪声的有效性。
In structured-light vision system, the accuracy of laser stripe center position is the main factor that determines the final system accuracy. The noise on the laser stripe is not removed in current extraction algorithms. Based on the analysis of these methods a novel extraction algorithm implemented by iterative fitting of B-spline was proposed. The pixels on the laser stripe were taken as control vertices for B-spline fitting. The noise was iteratively replaced by the pixel on the B-spline at the same horizontal position. A gravity model approach was applied to extract the laser stripe center after the noise removed. The number and the distribution of the pixels used in this algorithm were analyzed with the emphasis. The extraction accuracy of this algorithm was proved higher than others via contrast experiments. This demonstrates that the noise is removed effectively.
出处
《光学技术》
EI
CAS
CSCD
北大核心
2005年第3期430-433,共4页
Optical Technique
关键词
结构光
B样条曲线
光条中心
迭代算法
噪声
structured_light
B_spline curve
laser stripe center
iteration algorithm
noise
