Casting identification character recognition method based on CRNN-CTC

Abstract

The invention provides a casting identification character recognition method based on CRNN-CTC. The method comprises the following steps: collecting pictures containing casting identification characters to construct a data set; carrying out data augmentation on a data set, and solving the problem of small number of pictures by utilizing a method of rotating, adding noise and adjusting brightness and contrast so as to enhance the robustness of the model; preprocessing a data set picture, inputting the preprocessed data set picture into the established network model, carrying out feature extraction on the picture through a CNN, then outputting features to an RNN, and carrying out transcription through CTC; and calculating the loss function of the CTC, continuously optimizing the network model through back propagation, and ending the training until the best prediction effect is achieved. According to the OCR recognition method combining CRNN and CTC in deep learning, the recognition accuracy can reach 98.8%, and the model obtained through training has good generalization ability and fault-tolerant ability.

Description

technical field [0001] The invention relates to the technical field of image character recognition, in particular to a CRNN-CTC-based method for character recognition of casting identification characters. Background technique [0002] In recent years, with the development and popularization of computer technology, the level of modern technology has also increased rapidly, and image processing technology based on computer technology has risen rapidly and has been widely used in various fields. Today, image processing technology is almost closely related to people's lives, and image recognition technology has greatly brought convenience to our lives, reducing the labor intensity of staff in multiple industrial production processes and reducing the cost of industrial production processes. Error rate, while greatly improving production efficiency. [0003] With the rise and development of deep learning, computer vision has undergone tremendous changes and transformations. As a...

AI Technical Summary

Problems solved by technology

[0004] Since the characters on most castings are concave-convex characters on the surface of the object, the characters are the same color as the parent body and are three-dimensional characters, most of the mature products that recognize the characters are only for a single product, and the specificity is too strong

Although there are many mature character recognition products on the market, any one has limitations and cannot be applied to multiple scenarios

Images