The invention provides a piston ring closed-mouth gap and roundness zero-damage measurement method based on computer vision according to the computer vision theory, the image processing theory and mathematical knowledge and particularly relates to a piston ring detection method based on monocular vision detection. The method includes the steps of image enhancement, edge detection, roundness measurement, local sub-pixel processing, gap measurement and the like. According to the method, local image processing is conducted on a gap portion through the sub-pixel technology, so that the measurement accuracy is improved; the length of the gap is measured, and the roundness of a piston ring is measured, so the measurement range is expanded, and the detection quality is improved.