Recognizing the Mass of Parts in Drawings Using Neural Network Models

The article discusses the application of modern artificial intelligence methods, in particular neural networks, to solve problems in mechanical engineering. This area is becoming increasingly important, as process automation can significantly increase production efficiency, reduce costs and improve product quality. The paper proposes a method for automatic finding and recognizing the mass of a part in drawings based on the use of neural networks adapted to perform the task. To implement this approach, a new model based on the YOLOv5 architecture has been developed, trained to analyze drawings and allowing them to identify the area containing the mass of the part accurately. Next, a program was developed in the Python programming language, which uses the resulting model and the EasyOCR optical character recognition library to automatically extract information from drawings containing the mass value of the part, and then save this data to a Microsoft Excel spreadsheet. The program has been tested, and the results of the analysis have confirmed its high efficiency. It is able to recognize a text accurately, which allows you to process large amounts of data quickly and accurately. This opens up broad prospects for its application in various areas of technological preparation of production: to carry out automated selection of equipment, fixtures and means of production necessary for the manufacture of parts, calculation of material consumption, and grouping of parts. Thus, the developed program is a powerful tool for automating routine tasks related to the analysis of drawings, and can be widely used in various industries.