The Instrument Dial Numeric Characters Automatic Identification Technology

Digital instruments have a wide range of applications in the fields of industrial and testing with its advantages of high precision, being easy to read and to set up. Such as the testing of instruments, the acquiring of scientific test results and control system data, the debugging of the system and so on. With the development of the science and technology, the detection means of various sectors is undergoing the gradual transition from manual inspection to automatic or semi-automatic mode. Especially in the chemical and metallurgical industries, as well as in the harsh environment where the real-time detection is needed, in order to improve production efficiency and reduce misjudgment caused by human factors, auto-detection system can be used to complete the control tasks. To test automation, we first need to collect real-time, high-speed display parameters from the digital instrument. However, some instruments do not provide interface data communication with the computer. To achieve high-speed test automation, we need to use camera technology, image processing, image recognition technology and communications technology. Therefore, the study on digital instrument automatic identification shows a wide application.This paper will identify the number of objective accurately and quickly. A pretreatment is taken on the source images under natural conditions obtained. And then the study identifies the detailed target figures to get the method of finding holes and the edge-point method. Compared with other digital identification methods, this method only deal with the structural feature extraction to image recognition and classifies them without the need for charactering size and type of standardization,, and then get purpose of the identification. In addition, this article considers deterioration of characters caused by noise and mechanical meter shaft rotation, to make recognition accuracy enhanced.The main technology used is described as follows.(1) Image preprocessingIn this paper, the pretreatment process is given as follows: Graying - Filter - Binarization - Correction - Split -Thining, A new approach is proposed base on the existing research to deal with binarization, correction, segmentation.(2) Character features are extracted Holes and gaps are very important structural features to the numbers. Figures 0,4,6,8,9 have obvious holes,and 2,3,5,7 have obvious gaps. In judging these figures, the use of such structural features is a fast and effective method.(3) Check the resultsEach of the identified characters will have a check and this will enhance their accuracy.