Research On Key Technology Of Precision Measurement System For Geometrical Sizes Of Chip Components

Measurement technology is an integral part in the development of industrialproduction. In the present, as the inevitable product with the development of electronictechnology, chip components was moving towards lightweight, miniaturization, highreliability, and it played a very important role in the high level of integration circuit.However, human detected the feature sizes, which still stayed on the traditionalmeasurement methods, such as vernier caliper, micrometer, and tool microscope wereused for artificial sampling inspection. These methods had no advantage when it cameto the measurement precision, efficiency, the intensity of labor, and degree ofautomation and so on. Machine vision has the advantages of flexible, high precision,rapidity and high automation, which can detect parts efficiently and continuously, and itis a good way to satisfy the detection requirements for chip components.Aiming at the existing size measurement limitations of chip components, this papercombined with the structure and function of typical visual system, and researchedprecision measurement system for geometrical sizes of chip components by using thevirtual instrument technology and digital image processing technology. It includedoverall structure design of the system, hardware configuration and its type selection,implementation of the control unit, the analysis of image processing method, systemcalibration experiment, etc. Generally speaking, the key research content was consistedof three parts. Firstly, by means of two-dimensional plan and three-dimensionalstructure, the paper designed mechanical structure of the three-dimensional servoplatform, chose components such as the light source, lighting way, lens, camera, and thehardware platform of measurement system was constructed finally. Secondly,completing motion control of three-dimensional servo platform by LabVIEW, the papersolved the problems of programming and man-machine interface editing in the controlunit, and realized chip components positioning and auto-focusing technology of cameraon this basis. Thirdly, analyzing the principle and workflow of image acquisition, thepaper completed the task of image acquisition of chip components. This paperexpounded methods of image processing and analysis, meanwhile, compared imageprocessing algorithms to choose appropriate means for image processing system. Itincluded histogram equalization to enhance images, median filtering for noise reduction, through the threshold segmentation for image binarization, canny algorithm for imageedge detecting and linear fitting by the least square method.The paper took the traditional calibration method to carry on the experiment studyfor the system finally, which can calculate the pixel equivalent for transfer pixels tophysical sizes. And taking type of0805resistor as objects, we do the experimentalresearch by the vernier caliper and the built measurement system, repeated10times.The system measured the average of length and width was2.0876mm and1.2753mmrespectively. The experiments showed that the system has high measurement accuracyand measuring result within its size deviation. Comparing the measurement results ofthis system and vernier caliper, we drew relative error of length and width was1.32%and5.91%respectively, which suggested that hardware design and software analysis ofthe system was reasonable and feasible, that is to say, the measuring system that thepaper studied could realize the purpose for measuring geometry size of chipcomponents.