Image measurement is a method to obtain parameters, such as size, position, quantity and outline. It is performed using digital image processing technique to analyze and study the image of the object according to the computer vision principle.In this paper image process, analysis algorithm and other related technology of image measurement are discussed. Its main contents are as follow:1. Review the development and application in size measurement of computer vision and introduce computer vision theories, image measurement principle and general structure of image measurement system. To measure big size sheet part with holes automatically, a kind of automatic image measurement system with precise x-y platform is introduced, which can perform visual guidance localization and automatically taking pictures. In this system two cameras are applied, G for overall part and L for details of the part. By analyzing overall image camera L is guided to automatically shoot part details which need to be measured. What is drawn from these images by image processing and coordinate parameters feedbacked by x-y platform together contribute to reach results.2. Study and discuss the core part of image measurement system - algorithms and related technology. According to the general flow of image processing algorithms, such as pretreatment, edge detection, image segment, object detection, parameter measurement, sub-pixel and related technology are introduced and analyzed. Also, two methods are proposed, one is for auto-adaptive circles detection, the other for sub-pixel. While being discussed, these algorithms are realized and corresponding effect is displayed.3. Experiments and error analysis. In this paper several measurement experiments and error analysises were made to test sub-pixel algorithm and to simulate image measurement system and primary conclusions are reached. |