The Method Of Collecting And Processing Of Image Of Gauge

As one of the instruments used in manufacturing process and scientific experimentation, gauge is generally used in measuring and controlling. The usual gauge includes marine gauge, mobile gauge and aerial gauge and so on. People use the gauges to get the parameters. In some situation it is inconvenient to get the parameters from the instruments directly, but only from the gauges' reading for integrative process. The auto collecting, analyzing, processing and recognizing of the images of gauge can be accomplished by adopting the design thinking based on computer vision technology, and then providing the parameters for system surveillance.This paper studies the collecting and processing of the image of gauge. First, the collecting method and the makeup of the whole system are studied. Then the pretreatment process is studied in emphasis, including gray scale normalization, picture evening, image segmentation, mathematical morphology transfer, line pick-up, line recovering, pick-up of printing numbers, and a fast maximum entropy thresholding algorithm is proposed in image segmentation. After that, the normalization of printing numbers and the pick-up of the characteristics are discussed, after the introduction of artificial neural network, a simple algorithm based on artificial neural network is proposed with the purpose of recognition of printing numbers. At last, a demo system for image collecting and processing of gauge is realized based on correlational algorithms.The experiments, in the case of using compass, show that the system has both good effectiveness on processing of image of gauge and high precision on recognizing readings. So the system is applicable for the situation that is inconvenient to get the parameters directly. At the end this paper analyzes the error of the experiments andproposed the imagination of improving the system.