It is important to check the position of the endotracheal tube on the portable chest radiography for patients in the intensive care units. The position of carina is not easy to identify on portable chest radiography due to the low image contrast and abundant noise at this region of interest. In this paper, a novel method is proposed to identify the position of carina. The proposed method is integrated with the rule-based image segmentation, contrast enhancement, selective thresholding, and morphological image processing. Experimental results show that the proposed method is robust and the success rate is 85.9%. It can be used to enhance patient safety by early detection and prompt correction of improper position of the endotracheal tube.