Research On Instrument Marking Of Engineering Drawing Based On Image Recognition

The P&ID diagram describes the flow and control of the refrigeration system.It is the basic design drawing showing the overall overview,and embodies all the technical requirements of the refrigeration system.The basis for the electronic control design of the refrigeration system is also directly derived from the P&ID diagram.it is required to accurately identify the electronic control instruments in the P&ID diagram during the design process of electrical control,providing a basic guarantee for the correct design of the electronic control system.Limited to technical means,the identification of P&ID diagram has always relied on engineers to do it manually.However,due to the complexity of refrigeration systems and control systems,engineers are increasingly challenged to identify electronic control instruments.In order to ensure the correctness and reliability of the electronic control system design,engineers spend more and more time on the identification of electronic control instruments,and the cost of electronic control instrument identification in electronic control design is also higher and higher.In the early days of computer image processing technology,engineers have tried to apply computer image recognition technology to the identification of engineering drawings.However,due to hardware and algorithm limitation,image recognition is unsatisfactory.With the rapid development of deep learning technology,especially the application of convolutional neural networks in image recognition,modern image recognition systems have been gradually applied in production practices All of these make it possible to apply the image recognition system to identify the electronic control instrument.In order to reduce the cost of the electronic control system and improve the work efficiency,the electronic control instrument in the P&ID diagram has the characteristics of unified abstract structure and standard icon.The thesis is based on the modern image recognition technology,and the automatic labeling of the electronic control instrument in P&ID is carried out.the study.Through the research of this project,the possibility of automation of instrument marking is discussed,and the economic significance and technical significance brought by the application of this automation system to the electronic control design of the refrigeration system are discussed.This thesis describes the development of modern image recognition technology,the technical characteristics of fully connected neural networks and convolutional neural networks and its application in automatic image recognition.By discussing the characteristics of instrument icons and the requirements of instrument marking in P&ID diagrams,the advantages and disadvantages of fully connected neural networks and convolutional neural networks are compared.The requirements for image recognition technology for instrument icon recognition are considered,and the accuracy and operation cost are considered.In many aspects,this thesis constructs an automatic image recognition system based on convolutional neural network.By using the instrument icon and instrument number characters to train the image recognition network,the image recognition network is successfully identified in the electronic control in the P&ID diagram.The instrument is marked and the concept of automatic identification of the electronic control instrument is completed.This thesis introduces the latest computer technology achievements into the traditional engineering design field,greatly improved the efficiency and automation level of engineering design,reducing the cost of system design and development,and improved the overall computer application level of the company.At the same time,according to the characteristics of instrument identification,this thesis makes targeted improvement and adjustment to the identification network,enhances the training data,reduces the training complexity,improves the recognition efficiency and accuracy,and makes the accuracy rate and the leak recognition rate exceed Artificial identification.All of these proves that the application of image recognition based on convolutional neural network has great advantages in engineering drawing recognition.The image recognition system based on convolutional neural network is applied to the electronic control instrument identification in the P&ID diagram.The experimental results show that compared with the early image recognition system and manual recognition,the convolutional neural network significantly improves the efficiency of instrument identification and reduces the identification time,the technical development cycle and design cost of refrigeration equipment,improve the accuracy of drawing design,save engineers' design time and design load,and are typical applications of new technology in product development and design.The research have good technical and economic benefits and have the potential and broad application prospects of other design links.