Research And Implementation Of Coding Character Recognition And Temperature Measurement System For Hot Strip Rolling Billet

In the process of promoting the transformation and upgrading of the manufacturing industry,the steel industry is a typical example of the traditional manufacturing industry.The contradiction between its good information collection foundation and the backward manual management model needs to be resolved.Introducing automated and intelligent software models and gradually reforming them has become a breakthrough point for the steel industry to promote intelligent manufacturing.To realize the automation of billet charging statistics and automatic measurement of billet temperature in the hot continuous rolling production line,the system of code character recognition and temperature measurement for billet was developed.The system is used to replace the lagging management method of manual statistics of billet entry information in the existing model,improve the efficiency of raw material management,and achieve refined management of billet charging and control scheduling.The main work and innovation points of this paper mainly include the following contents.(1)The thesis takes hot-rolled billet coded character recognition as the research object,and has conducted a lot of theoretical research on traditional character recognition methods.Through experiments,the image processing algorithms of each link of traditional character recognition have been analyzed and selected.The application of BP algorithm in billet character recognition is optimized by comparing the template matching method and BP network algorithm.On this basis,the Halcon image processing operator was used to achieve the initial rapid construction of the system.(2)To improve the accuracy of billet code recognition,the target detection method based on deep learning is studied.Combining the features of the Faster R-CNN series based on candidate boxes and the characteristics of the end-to-end YOLO method,the SSD network model is applied and improved.Using TextBoxes’ character recognition ideas,the recognition accuracy of the billet code is improved,and the real-time performance is high,which meets the requirements of the billet automatic statistics in the hot rolling furnace area.(3)In the image acquisition and temperature detection design,it innovatively combines the two,flexibly uses the dynamic detection function of camera.By monitoring the picture file and using socket communication method,the billet temperature is automatically measured,which is validated by field application.(4)Combined with the production process,the design of temperature measurement method and the highly accurate deep learning character recognition method,the on-line recognition of coding characters and the rapid temperature detection of hot rolling billet are realized through the selection design and construction of hardware and software.The billet coding recognition and temperature measurement results are of great significance for the real-time production tracking of MES system and the corresponding real-time generation of laser coding information,which promotes the intelligent transformation of hot continuous rolling production line.