Grid Row Parameter Identification Using Tensorflow Convolutional Neural Network

Text recognition is an important field of pattern recognition application. In this paper, a joint parameter recognition method based on convolutional neural network is proposed to solve the problems of small font, various cables, and traditional manual detection difficulty and low efficiency. Based on TensorFlow framework, this method builds a convolutional neural network model that can realize end-to-end recognition. First manually adjust the shade number serious cables, from different angles to adjust good cable collection into image preprocessing, and then use the contour detection and projection image segmentation algorithm on area of the wire number to find out and cut them into a single character, after normalization processing packaged into the sample set and test set in proportion to the convolutional neural network training, Finally, the obtained joint parameters are compared with the standard wiring information base, and the wrong or missed cables are screened out for manual adjustment. The convolutional neural network model is applied to power grid cable identification, and the results show that it can accurately locate and identify the routing parameters, effectively assist manual detection to identify the cable parameters, reduce the error rate of human eye detection, and greatly improve the efficiency of manual detection.