Study On The Embossed Character Recognition System Based On Machine Vision

Tire embossed character is widely used in the industry of a class of identification characters,it is the use of molds to make the surface of the surface deformation and embossed.The character of the embossed is the same as the color of the background area,and it is a solid character.With the improvement of modern manufacturing technology and production efficiency,automatic recognition of embossed characters has become an inevitable requirement of intelligent production management of industrial products.Because there is no color difference,it is difficult for the industrial camera to acquire the characters of the imprint image,so it is difficult to borrow the existing optical character recognition directly.This paper aims at the problem of hard to recognize the characters of concave convex characters.The following research work has been carried out.An image acquisition hardware system for tire embossed characters is designed.This paper first describes the composition of the machine vision system,analyzes the performance parameters and selection methods of the related hardware,and selects the appropriate hardware according to the specific requirements of the visual system of this topic.Analysis of the effect of light on the imprint of character image,and according to the lighting scheme selection of high angle annular light source LED front lighting characters as the best tire imprint lighting scheme recognition system,for the follow-up processing algorithm has laid a good foundation.The template matching method and the edge detection and segmentation method are designed for the ring arrangement of tire character information.This paper first introduces the characters of common positioning algorithms,and then use the embossed characters and the known character information based on the normalized cross correlation(NCC)problem of the pressed character location,then known information is proposed based on rough positioning method,positioning the tire sector scanning character area domain,better location,can accurately determine the character of tire the region boundary.By comparing the experimental results of several segmentation methods,the edge detection is used to segment the character images that have been positioned and converted into horizontally arranged characters.The method of character recognition based on least squares support vector machine is adopted.This method solves the problem that the training time of the embossed characters is long and the accuracy of recognition is unstable.From the final software to achieve the effect of view,this design of the embossed character recognition system to identify the correct rate and stability in line with the design requirements of the system,to achieve a good recognition effect.