平顶线结构光的中心检测算法及光刀平面标定

发布时间：2018-10-09 12:10

【摘要】：针对传统线结构光光刀平面标定方法测量精度不高,应用范围小的问题,提出基于平面标靶的线结构光系统光刀平面标定,对无激光的标靶图片进行迭代摄像机标定,有激光的标靶图片进行光刀平面标定.提出光强符合均匀分布的平顶激光检测中心算法,将平顶激光建模为矩形的台阶函数,估计背景亮度和前景亮度,确定亮条纹宽度,再将窗口内的有效像素参与重心计算,得到光条纹中心.用该算法对不同噪声及不同量块的图片进行处理,结果表明,处理后图像的均方根误差分别在0.149pixel和0.176pixel内,表明该算法抗噪声能力强、精度高.用该算法提取光条中心,计算光条在标靶上的位置,根据至少两个姿态下的光条中心三维点,基于最小二乘法拟合光刀平面.通过迭代摄像机标定和光刀平面标定,利用三角测量法,在立体视觉模型下获取物体的三维点云数据.实验测量两个距离为100.5mm的标准球,相机与标准球距离为500mm,比较两球心距离与标准距离,测得平均误差为0.236mm.表明平顶激光检测中心算法切实可行,光刀平面标定方法基本满足要求.
[Abstract]:Aiming at the problem that the measurement accuracy is not high and the range of application is small, the plane calibration of line structured optical knife based on plane target is proposed, and the target image without laser is calibrated by iterative camera. The target images with laser are calibrated. A flat-top laser detection center algorithm with uniform light intensity distribution is proposed. The flat-top laser is modeled as a rectangular step function to estimate the background brightness and foreground brightness, to determine the width of the bright fringe, and then to take the effective pixels in the window into the calculation of the center of gravity. The center of the light stripe is obtained. The results show that the root mean square error of the processed image is in 0.149pixel and 0.176pixel respectively, which shows that the algorithm has strong anti-noise ability and high precision. The center of the strip is extracted by this algorithm, and the position of the strip on the target is calculated. According to the three dimensional points of the center of the strip under at least two postures, the optical knife plane is fitted based on the least square method. Through iterative camera calibration and optical knife plane calibration, triangulation method is used to obtain 3D point cloud data of objects under stereo vision model. The distance between the camera and the standard ball is 500mm. The average error is 0.236mm. the average error is 0.236mm. It shows that the algorithm of flat-top laser detection center is feasible and the calibration method of optical knife plane basically meets the requirements.
【作者单位】：上海大学机电工程与自动化学院;机械系统与振动国家重点实验室;
【基金】：国家自然科学基金(Nos.51575332,61673252) 国家重点研发计划子课题(No.2016YFC0302401)资助~~
【分类号】：TP391.41

【相似文献】