The invention discloses a ceramic membrane surface defect two-dimensional detection method realized on the basis of a machine vision technology, and adopts a digital image processing technology to construct a ceramic membrane surface defect detection system consisting of a mechanical unit, a detection operation unit, an image collection unit, an image processing unit and a result display unit. A median filtering operator is adopted to preprocess a surface image; the extraction effects of different edge detection operators for a ceramic membrane edge are analyzed, and a Kirsch operator and a Canny operator are selected to be combined to realize the edge detection of the surface defect; an iterative threshold value segmentation method is selected to serve as a spare algorithm of a subsequentprocessing process; and after the area and the length-width ratio of a two-value defect image are calculated, a feature threshold value index is adopted to judge whether the ceramic membrane is a qualified product or not to realize the classification of quality products, inferior-quality products and defect types. By use of the method, the ceramic membrane surface defect two-dimensional detectionrealized on the basis of the machine vision technology can be realized, a detection method is quick, intelligent, safe and reliable, and detection efficiency is improved.