By combining true roundness algorithm and machine vision technology witch includes image acquisition, image preprocessing and camera calibration，developing a morphological analysis method based on image processing and an edge fitting method based on least square method (LSM), monitoring the hole making of irontower components processed on punching machine tool whose size, number and position are not fixed, and evaluates the monitoring results and practicability of the two methods. The average radius accuracy of edge fitting method could be reached at 0.395 mm, and the accuracy of position detection could be reached at 0.399 mm, which could meet the requirements of both the radius of hole making and the center position of the inspection component, realize automatic monitoring of products online and replace the previous image stitching matching method and motor driving control block by software programming and auxiliary algorithm. Automatically collect images of tower components entering steel joint production line conveying area, and provide technical support for building visual inspection system of machine tool punching.