The Research On Technology Of Recognizing Characters In Power Equipment

This article has researched on the characters recognition of equipment scutcheon in the electrical power system, studied and achieved the every module of the scutcheon character recognition system. Explanation has been carried out from four aspects, such as image preprocessing, scutcheon location, binaryzation algorithms of scutcheon, and the character recognition. A new binaryzation algorithms adapted to scutcheon has proposed after analysis the features of equipment scutcheon which can extract the steel seal information on the scutcheon. It has preprocessed the equipment image because it is inevitable for noise pollution. The preprocessing method contained de-noising and gray level correction which can extrude the interesting fraction. Doing edge detection of scutcheon by five edge detector, such as Roberts detector, sobel detector, prewitt detector, LoG detector and canny detector. Extracting the scutcheon frame by the method based on Hough transform, analysis the factors affected the correctness and proposing an ideal scutcheon. The characters have been recognized by adopting OCR and the results of measurements have been analyzed. In the end of this paper, some expectations to the further research on this topic are put forward.