使用双传感器系统降低光刀图像缺损误差

Reducing Error Caused by Defaults of Light Knife Image by Double Sensor System

用线结构光扫描测量三维物体的过程中,由于物体表面的凹凸不平、反射率的不连续变化以及遮挡造成获取的条纹像存在各种缺陷,导致最后得到的三维面形数据存在偏差。基于此,提出了一种采用双传感器,并将两个传感器图像映射到光刀平面上,然后再对其进行面内融合处理最后提取中心线的方法。对此方法进行了计算机模拟,并对牙齿进行了测量。实验结果表明该方法实现简单,在存在条纹缺陷的地方,精度可以明显提高,有效地降低了条纹缺陷导致的数据偏差。

The defaults of laser-stripe image caused by the shape discontinuity due to the illumination, the discontinuity of surface reflectance and the sensor occlusion, will lead to the errors of three-dimensional （3D） surface data, while measuring the shape of 3D object by line structure light scanning. A new effective data-fusion method is presented, which can eliminate the errors by using double sensors. First, the two sensors＇ images are mapped into the light knife plane; second, data fusion will also be made in that plane; finally, the center line of the image in the plane could be extracted. The simulation and experiment of one tooth show that the method is very simple, can produce a higher accuracy, and can reduce the errors of the data that introduced by the fault of light stripe.