Image sub-pixel recognition method for optical precise adjustment

Cavity adjustment is a critical step for manufacturing laser gyro. Location relation of the laser facula and the aperture images should be detected to provide control information for the movement of the cavity adjustment mechanism. According to the image features, a sequence image variance method is proposed to segment the laser facula image, and maximum variance between clusters method (Ostu algorithm) is adopted to acquire automatic threshold value to segment the aperture image. Use the edge detection of binary image to obtain pixel edge of the laser facula and the aperture images. And sub-pixel edge point information of the images is given by a parameter fitting method. Centre points coordinates of the laser facula and the aperture are achieved by fitting sub-pixel edge utilizing the least square ellipse fitting method. The experimental results of cavity adjustment verify the validity of image recognition method of the laser facula and the aperture.