With the application of image processing technology in the hydrophobic detection of composite insulators, the traditional manual detection method has been gradually replaced. In this paper, the characteristics of the waterdrop image are used to transform the original image into the HSV color model for preprocessing. The advantage of the contrast between the water droplet and the dry region in the S component is used to the missing part of the interior and edge of the waterdrop image. Filling is performed, and then two different labeled watershed transformations are performed on the preprocessed image. The experimental results show that the algorithm can effectively segment the water bead image, and the shape of the detected water bead is correct. The area error is within 5%, which provides a basis for determining the hydrophobic performance of the composite insulator.