Camera calibration aims at building up the mapping relationship between 3D world coordinate and 2D image coordinate. The traditional calibration board contains dozens of neatly arranged standard circles or grids. Calibration is completed by extracting the center coordinates of standard circles or grid corner coordinates. In a micro-object measurement system, the camera field of view is very small and camera cannot extract enough points from the calibration board. To solve this problem, a new method based on conic and lines is proposed. This method uses conic and lines equation instead of points to calibrate so that enough information can be extracted from calibration board even under small field of view. Both computer simulation and real data test show that the proposed method has higher accuracy and better robustness. In addition, the calibration board is a standard semicircle which is very easy to be made and easy to be used under small field of view.