Three-dimensional recovering based on variable focal length via Bezier networks

Abstract

An automatic technique to perform measurements of variable range via variable focal length is presented. This technique provides measurements of variable range by means of a Bezier network based on a laser line projection. The variable range measurements are performed by adjusting the camera position and the focal length. These parameters are automatically computed during the vision task by means of the line position. Also, the initial calibration of vision parameters is carried out via image processing of the laser line. Thus, the proposed technique avoids external re-calibration, which is applied by the traditional methods of flexible measurements. This leads to an improvement of accuracy of the measurements of variable range via Bezier network. It is because the proposed method does not add errors of external re-calibration to the measurement process. This technique performs measurements of variable range via magnification of the surface depth, when it is not retrieved by the initial configuration. Thus, the system adapts the setup to the resolution of complex surface, which is being profiled. Furthermore, the adjustment does not interrupt the vision procedure because the change of vision parameters is calculated in a fraction of second. The contribution of this technique is elucidated by an evaluation based on the traditional measurements of variable range. Also, the time to achieve the surface measurements is examined.