Abstract: Objective Designed a method that can automatically distinguish the X-ray radiological positions based on image information. Methods For any selected input clinical X-ray digital images, we extract the image outline in the edge detection method using optimal threshold selection, then process binarization. After compared with the whole body phantom X-ray images, then process gray level similarity traverse detection. Results By optimizing the decision, the best matching region of the same input film was obtained, and the imaging part of the patient is determined based on prior knowledge. Conclusion The algorithm is fast and has a high degree of robust.