Double-contact constraint-based calibration method for a line-structured light vision sensor using a sphere target

This paper proposes a method to calibrate a line-structured light vision sensor utilizing a sphere target. A sphere target with a known radius is placed in two or more different positions within the measurement field of the sensor. Images of the sphere target with a light stripe are captured sharply by the camera in the sensor. Feature points of the sphere outline are extracted to fit the related projected ellipse on the image plane. Once the light stripe is determined, the vanishing line together with the normal vector of the light plane can be calculated based on geometric properties. By leveraging the spatial characteristics of the sphere target and image feature constraints, three-dimensional coordinates of double-contact points can be obtained precisely according to an iterative optimization algorithm, where the Levenberg-Marquardt algorithm minimizes the geometric distance between detected points and the intersection of the light plane and sphere surface. In this case, the complete expression of the light plane is confirmed. Both simulations and physical experiments are conducted to evaluate our proposed method. Experimental results indicate that the precision of our calibration is 0.014 mm/0.017 mm within a measurement area of approximately 230 × 180 mm. As an isotropy of a sphere and double-contact constraint, extractions of features and calibration results are precise enough, and perspective projection errors are eliminated. Furthermore, if the sphere target is positioned in more than three different locations, the intrinsic parameters of the camera can be determined synchronously.