In order to study blue ear epidemic early warning and monitoring method in large-scale pig farms, the non-contact ear automatic color detection method was proposed. Pig’s thermal infrared images and visual images were collected at the same time and the same viewing angle. Two pictures were used together to find the optimal scale factor of matching. By the optimal scale factor, pig ear root section can be found in visible image, and then the ear root central point can be confirmed. According to active shape model method, pig ear root central point was selected as the first feature point, the ear tip as 18th feature points, and both middle point of outlines as 9th and 26th feature points, and 34 pig ears outline feature points in all were selected by human-computer interaction. ASM search scope was defined in pig head region, thus pig ear outline could be extracted correctly. Then the extracted pig ear color was compared with the color of normal pig ears, the ear color detection accuracy could be above 77%, and it could be easily found whether there was the risk of blueear pig disease. The results showed that due to the limited search scope method, the pig ear contour could be extracted accurately, and it could be applied to auto ear color detection in swine house.