This article proposes a method of rack measurement based on computer vision. It establishes a computer vision measurement system; the system consists of precision linear guide, camera, computer and other several parts. The entire system can be divided into displacement platform design system and image acquisition system two parts. The displacement platform system is that the linear guide campaigns driven by the driver controlled by the computer, to expand the scope of this measure realizing the measurement for the whole tooth. Image acquisition system is the use of computer vision technology to analysis and identification the capture images, the light source emitting light to the caliper rack, camerawork is to be the image which acquisitioned. Then input the images to the computer through the USB interface in order to the image analysis, such as Edge Detection, Feature Extraction and so on. And the detection accuracy reaches to sub-pixel level. Experiment with the rack modulus of 0.19894 instruction sheet calipers to measure, using image processing technology to realize the edge detection, and getting the edge of rack. Finally get the basic parameters of the rack such as p and s, and calculated individual circular pitch deviation fpt, total cumulative pitch deviation Fp, tooth thickness deviation fsn. Then comparison the measurement results with the Accretech S1910DX3. It turned out that the accuracy of this method can meet the requirements for the measurement of such rack. And the measurement method is simple and practical, providing technical support for the rack online testing.© (2010) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.