A Two-Step Calibration Method for Vision Measurement With Large Field of View

Abstract

Nowadays, the commonly used large view calibration methods usually have the problems of low measurement accuracy, big on-site operation limitation, or high tolerance requirement for target preparation. In order to resolve the previously mentioned issues, a two-step calibration method for the interior and exterior parameters of multicameras is proposed based on the principle of industrial close-range photogrammetry. In the first step, by using a small-scale cross target at the closer range, the interior parameter of the front section of the camera is calibrated based on the pyramid method. In the second step, by setting a proper number of serialized targets at the further range of the measured area, the exterior parameter is calibrated based on space resection with multi-image. Finally, the interior and exterior parameters are all optimized by using the bundle adjustment method to get accurate measurement results. A number of experiments are performed to verify the feasibility and accuracy of this proposed method. As per the experiment results, with a camera resolution of <inline-formula> <tex-math notation="LaTeX">$2448 \times 2048$ </tex-math></inline-formula> pixels, the reprojection error is less than 0.11 pixels. For the measuring volume of <inline-formula> <tex-math notation="LaTeX">$10.0 \times 8.4 \times 3.6$ </tex-math></inline-formula> m, the maximum absolute error of 3-D measurement is 0.42 mm. The measurement error for the flapping angle of rotor whose diameter is 10 m is less than 4.5&#x0025;. By using this method, the interior parameter can be collected in the laboratory, while the exterior parameter can be collected in the open field separately. This method has high practicality under the actual engineering working environment.