Image Recongnition In Gauge Block Calibration

With the rapid development of science and technology，automation is widely used inmanufacture, which become an important means to efficiency. Metrology, one of theimportant science for promoting the productivity, also need to keep up with the pace of socialdevelopment, lead into new technologies in order to better serve the community and theenterprise.The vertical interferometer meets the equipment of precision, however, is not convenientto use. It works in low efficiency, and can not fix to the increasingly tasks. Therefore, we setup the research on block gauge automatic testing equipment. Now we have succeed in doingsimple measurements. This paper combined the numeral recognition and centering the blockgauges with the operating system of the verification device, intends to improve theautomaticity.The object recognition of the system are the numbers marked on the block gauges. Thispaper researched some theories in this field, including image enhancement, image orientationand segmentation, feature extraction and classifier design. In addition, main technologyfeature like feature extraction and classifier design is specific narrated in this paper.In orderto improve the classification accuracy, BP Neural Network is applied to it.The numeral recognition system is designed by Matlab, and rapidly builds the neuralnetwork and graphical interface by using the toolbox. When the numeral recognitionprogram is designed, the results meets the requirements through series test.In the finial part of this paper，introduces the reformed block gauge automatic testingequipment, from the experimental results we can see that the accuracy has been improved.