An improved camera calibration method for 3D flame measurements based on tomographic reconstruction

Camera calibration is a key process in tomography based measurement techniques, which determines the locations and orientations of cameras relative to the target. The state-of-the-art camera calibration method uses a 2D planar pattern (i.e., a chessboard) to register the angle and distance of cameras. However, the out-of-focus blurring happens when the camera observes the 2D pattern from a large inclined angle, resulting in big error (i.e., > 5⁰⁾ for the calibration and the consequent tomographic reconstruction. Therefore, this work reports an improved camera calibration method by using a 3D pattern instead of the 2D pattern. Based on the calibration method, the tomographic reconstructions were performed to resolve the 3D structure of a laminar diffusion flame. The reconstructed flame height and conical angle were compared with the theoretical results to validate the 3D flame measurement technique.