How to calibrate cameras rapidly and precisely is always a pressing problem in the visual measurement. Now there are many calibration mehods for different camer imaging models. A sorting algorithm for chess-board nodes based on the round marker, polar radius and polar angle is proposed, so that realize fully automatic calibration for single and binocular camera high-precision automatic calibration. Experiments vevify the effective and robust of the algorithm for different pose calibration images. Results show that the improved calibration ensures the calibration accuracy of the camera and improves the level of automation calibration, so it can be widely used in the calibration of single and binocular camera.