| Product marking and machine vision technology have been widely used in the industrial field,and the quality,reliability and production efficiency of industrial products have been greatly improved.In order to realize the full supervision of product production information,steel product manufacturers first mark the information on the surface of the product,then identify the mark and query the product information through the production information database.At present,manufacturers can only query the production information of bundles of special steel bars in the full supervision of special steel bar production information.The introduction of machine vision technology can realize the query of the production information of a single special steel bar and improve production efficiency.According to the shape characteristics of special steel bars and the complex production environment,a marking scheme based on the vertical double marking points on the bar end faces is designed.The direction of characters is limited by square and circular mark points,and the information mark image with random angle is rotated and corrected.Aiming at the image of the end face of the special steel bars with low contrast acquired by the image acquisition system,combining with wavelet transform,Gamma correction and median filtering algorithm,an image enhancement algorithm based on wavelet transform is proposed to enhance the contrast between characters and background and remove the noise in the image.The character region is detected and located by combining mser and Canny edge detection.On this basis,the double mark points are removed and the character region is divided into single characters based on projection method.Aiming at the problem of character recognition,SVM classifier is created,and a large number of character images are used to train SVM,so that the character recognition has a high accuracy.For the end faces image of the bundled special steel bars,the end faces image of the bundle of special steel bars is divided into a single special steel bars end faces image based on Hough transform.The recognition of the end character of the end face of the bundle of special steel bars is realized by combining the image enhanceme nt,character detection,character segmentation and character recognition algorithms.According to the production information management mode of special steel bars and the engineering practice of product end faces marking.,the information marking scheme of special steel bars is designed,an enhancement algorithm for low contrast end faces image of special steel bars is proposed.The characters in the end faces image of single special steel bar are detected,segmented and recognized,and the segmentation and character recognition of the end faces image of bundled special steel bars are completed.It is necessary to establish the technical foundation for realizing the traceability of the whole process of the production information of single special steel bars. |
PDF Full Text Request