Study On Camera Calibration And Single View Metrology

In this paper, the study is focused on camera calibration and single view metrology.Planar-pattern based camera calibration technique, which usually involves recognition and matching of model plane, plays an important role in all of camera calibration techniques. The traditional technique for camera calibration on a chessboard is usually accomplished by means of the human-computer interaction, which decrease efficiency of the algorithm. So an effective algorithm is proposed for automatic recognition and matching of checkerboard pattern. 1) A corner detector is designed to detect the internal chessboard corners; 2) Grid lines of the chessboard are obtained by using Hough transform under the constraint of vanishing point; 3) The recognition of model plane is accomplished based on the characteristics of chessboard corners and grid lines. The corresponding relationship between the checkerboard pattern and its image is built; 4) The intrinsic and extrinsic parameters of the camera is calculated. Experimental results show that the proposed algorithm performs excellent in the chessboard pattern. Compared with traditional ones, the algorithm obviously improves the speed of calibration process. It is especially adapted to automatic calibration based on multiple images.Recently, the image-based measurement technique has attracted a lot of attention in the field of computer vision. Using a single view image to measure world distance is a very important branch. A current method of the plane metrology is to measure the geometrical entities on a scene plane (referred as the reference plane) via the homography between the plane and its projected image. On these bases, this paper shows how the projection matrix is obtained by weak camera calibration if getting knowledge of the perpendicular vanishing point and how the perpendicular height is measured. It can be seen from both simulation data and real image experimental result that the proposed method has good measure accuracies and can be applied to structured scenes measurement.