We present an automatic method for region of interest (ROI) location in camera calibration used in computer vision inspection. An intelligent ROI location algorithm based on the Radon transform is developed to automate the calibration process. The algorithm remains robust even if the anchor target has a notable rotation angle in the target plane. This method functions well although the anchor target is not carefully positioned. Several improvement methods are studied to avoid the algorithm's huge time/space consumption problem. The algorithm runs about 100 times faster if these improvement methods are applied. Using this method fully automatic camera calibration is achieved without human interactive ROI specification. Experiments show that this algorithm can help to calibrate the intrinsic parameters of the zoom lens and the camera parameters quickly and automatically.