License plate identification system and method thereof

A license plate identification system, comprising:

• at least one server apparatus, comprising at least:
• a license plate data storage module, storing at least one license plate picture file to be identified and a plurality of license plate sample picture files, wherein each of the license plate sample picture files comprises a label area, and image content of the label area is an image character feature, the image character feature is a double-row character or a single-row character;
• a neural network learning module, connected to the license plate data storage module, used to perform deep learning training on the label area of the license plate sample picture files to generate a neural network model;
• a license plate identification module, connected to the license plate data storage module and the neural network learning module, used to input the license plate picture file to be identified into the neural network model for outputting an analysis result information;
• and
• a decoding output module, connected to the license plate identification module, used to perform decoding on the analysis result information through a decoding algorithm for obtaining an identification license plate character content;
• wherein the neural network model comprises at least a plurality of convolutional layers and a deconvolutional layer, wherein the plurality of convolutional layers image extract image content of the label area as a feature map for rearranging character features in the image content, and the deconvolution layer is used to amplify the feature map to increase upper limit of identifying string length, wherein the feature map corresponds to having a plurality of time feature regions;
• wherein the deconvolution layer is connected to a character feature extraction layer, the character feature extraction layer extracts a character feature matrix for each time feature region of the feature map based on a plurality of character types, and the character feature matrix at least comprises an output channel quantity information, a vertical direction feature information and a horizontal direction feature information, wherein the output channel quantity information is a character type quantity, wherein the vertical direction feature information is height of the time feature region, and wherein the horizontal direction feature information is width of the time feature region.