In view of the actual situation of low measurement efficiency,low precision and little metering data caused by manual sampling detection for steel pipe end wall thickness,the machine vision measurement technology was studied,the online detection system of the steel pipe wall thickness based on machine vision was designed.The relationship between the image pixel value and the actual value was obtained by calibrating the camera.The images of steel pipe end captured by the camera were processed to image preprocessing,edge feature point extraction and other operations.The RHT circle detection algorithm was improved to detect the round profile of the steel pipe end surface,and the wall thickness of the steel pipe was measured.In order to facilitate the operation,the human-machine interface software of steel pipe wall thickness detection system was designed based on Qt cross-platform development framework.The test result shows that the system is of good stability,with measuring error of about 0.1 mm,which can efficiently complete the wall thickness detection task.