Recognition Method And Realization Of LCd Aviation Instrument Information Based On Image Processing

With the continuous development of computer and electronic technology,video image processing technology has been widely applied in industrial production,medical diagnosis,resources,environment,meteorology,traffic detection and cultural education,and other fields and achieved great economic and social benefits.The utilizing of video image processing technology in military avionics field,such as study on the main instrument of aircraft cockpit display to realize the authomatic identification of the pilot operation pictures,however,is a wholly new research area that very few scholars have put their efforts in.This paper focuses on the LCD display images in a certain type of aircraft cockpit,and tries to detect and identify the information of instrument images by utilizing the image processing technology.According to the characteristics of virtual instrument LCD image,a set of video acquisition system has been constructed on PC platform to fulfill the real-time acquisition of the images,using the MFC software framework,the DirectShow technology and the video acquisition card supporting WDM mode.Functions,such as the preprocessing of images,extraction of image features,and detection of charaters,have become possible by third-party open-source of computer vision library such as OpenCV.Identification of charater information has also been realized byOCR optical character recognition technology.This paper has mainly studied on the image preprocessing algorithm and realized the color image grayscale and binarization transformation.Through the study of several kinds of commonly used template matching algorithm,the sum of squared residuals template matching algorithm,which is based on the reduced sampling,has been presented in this paper as well.Compared with the ordinary ones,this new algorithm has sevearl advantages such as simplification of calculation,speeding-up of the matching speed,and even the efficient aquisition of preprocessing images.Through the research on the angular point of images,this paper has raised a algorithm of preliminary extraction of charaters that is based on Harris' s denseness of angular point.Meanwile,with the help of the connected domain features of images,the complete elimination of interferentialinformation and characteral detection of instrumental images have become possible.In addition,the identification rate of instrumental images has been successfully increased by integrating the advantages of Tesseract-OCR and MODI,two different algorithms of identification.According toa comparative study toward those two algorithms,it has been found that the former one advantages in efficient and effective identification of numerical information while the latter is adept at accurate identification of characters and letters.To conclude,this paper has provided a case that instrument image recognition system can effectively achieve the detection and identification of information on the instruments such as numbers and characters.What is even better,this system has currently been successfully applied to the authomatic test in a certain type of aircraft.