An adaptive Canny operator edge detection method for detecting the edges of infrared images of a ship is proposed. In the method, the maximum local standard deviation of an image is used to set Gaussian filter scale parameters for the reference value. Then, the maximum variance method based on a gradient histogram is used to calculate the high and low thresholds of the Canny operator so as to implement the automatic detection of edges. The experimental result shows that the ship edges detected by this method are clear and continuous with less noise. The method is better than the traditional Canny operator edge detection method and has a good engineering applicability.