The Recognition Of Digital Indicator In High-voltage Transformer Substation

The recognition of digital indicator in high-voltage transformer substation is one typical application of robot vision. The automatic recognition of digital indicator may enhance the working ability and environment adaptiveness of mobile robot in certain degree, and may greatly enhance the working efficiency of the system. The automatic recognition is also helpful to realize nobody management of the transformer substation, to reduce the accidents formation rate.The images are gathered through a CCD camera mounted on the platform of a mobile robot which patrols in the transformer substation. The images are then preprocessed in order to enhance the clarity of the images. Medium filter, Laplas filter and binary processing are performed in the image preprocessing. For separating the object from the background, a suitable threshold value is set through the analysis of the grey scale graph.The features are obtained by choice and withdrawing methods. A multilayer classifier scheme is used to recognize digital readings. The character width feature and character formation trajectory feature are employed in coarse recognition .The net grid feature and intersection spot feature and graph connection feature are used for fine classification. This scheme enhances the recognition efficiency greatly. The single character recognition is realized by template matching. The computation of Euclidean distance between the training sample feature and the test sample feature is carried out to achieve recognition according to the similarity measurement principle.Experiments have conducted to verify the proposed methods. Hardware system including CCD camera and image gathering card is introduced .The image processing, feature extracting and recognition routings are programmed by using object-oriented program language Visual C++ 6.0 under Windows XP operating system environment. The experimental results have shown that this system can obtain the satisfying recognition results.