Abstract: In order to reduce the intensity of manual work and improve the detection speed and quality of the wavelength division multiplexing (WDM) component, a facula automatic detection device of WDM component was developed based on machine vision. A four-axis motion platform combined with a silicone vacuum nozzle was used to achieve automatic grasping and posture adjustment of WDM component. Charge-coupled device (CCD) camera was designed to acquire the image of WDM component. The contour and the deflection angle of the component were obtained through grayscale and edge detection. Moreover, complementary metal-oxide-semiconductor (CMOS) camera was used to obtain the exiting facula of WDM component. The center positions and quantities of the multiple faculae were determined by using Hough circle detection. Furthermore, the facula center points were fitted into a straight line through the least squares method, and the quality of the optical component was evaluated based on the slope of the line. The experiments demonstrate that the facula automatic detection device can complete the quality inspection of WDM component within 37 s, which is three times faster than manual inspection, and the inspection accuracy is higher than 97%. The research provides a reference for the development of a high-speed and high-precision facula automatic detection device of WDM component.