A Research On The Automatic Image Measurement System For SMT Stencil

With the rapid development of information technology, Surface Mount Technology, as an advanced assembly technology is applied quickly and widely in kinds of electronic products and assemblies. Not only SMT stencil manufactures need a high efficiency, high precision products measurement means, but also SMT processing factory need the measurement of the wearing of stencils in the printing process frequently. Machine-vision measurement technology will be an important means for ensuring stencil processing quality, as the target scale get smaller, the ever high production rates as well as the precision detection requirement.This paper is supported by the project"Research & Development on the Automatic Image Measurement System for SMT Stencil Based on the Machine Vision Technique"(2009B010900017), which is a science and technology plan project of Guangdong. Based on computer vision theory and image measurement technique, key technologies for Surface Mount Technology (SMT) stencil detection is researched in the paper, which contain image processing, path planning, and statistical process (SPC) and other implementing methods. An auto image measurement system for Surface Mount Technology (SMT) template is presented, which can measure the parameters such as hole roughness, hole size, hole shape and hole position of SMT stencils by automatically tracking the stencils'edges and evaluate stability of SMT process with Static Process Control (SPC).The Combination platform between the system's hardware and software is designed, in which three parts are involved: image collecting module, image processing module and motion control module. The image collecting module consists of a CCD camera, a measuring microscope and an image collecting card, which is used to complete images collection. The motion control module consists of a personal computer, a motion control card, grating and so on.A series of image processing method for SMT stencil image is presented in the paper, which is consisted by median filter, iterative threshold segmentation, Sobel operator Edge Detection to detect image edge. The roughness measurement of SMT stencil is measured. A graphic element Classified method and an improved nearest neighbor approach are proposed with which the system can optimize the path planning to move along the measured SMT stencil edge by the motion control system and the mechanical system.The existing problem is analyzed while using control chart and process capacity indication to evaluate stability of SMT process with necessary basic SPC theory. The size and position of SMT stencil detection parameters are sample measured, which are control by different control charts at the same time. The results meet the requirement of SMT stencil measurement.