Research And Implementation Of Instrument Recognition System By Video Scanning

Based on the research of the structure of instrument recognition system, image preprocessing technology, Hough transform algorithm and neural network,this thesis designs the flow of image preprocessing, realizes the image preprocessing algorithm,and improves the Hough transform algorithm and BP algorithm.The instrument recognition system made by video scanning has realized the reading recognition of index-instrument and digital display instrument. The image of instrument plate is obtained by a digital camera firstly, then an image preprocessing is carried through. For the index-instrument,the system adopts the Hough transform to recognize the pointer's rotating angle,then get the reading according to the relation between the reading and the pointer's rotating angle.For the digital display instrument, character features are extracted by 15-feature extraction method, and the reading is recognized by the BP neural network. The experiment result proves that the system is a user-friendly system, the recognition is of high veracity and great accuracy,with the value of practical application.