Automatic Maximum Diameter Determination of Rotation Parts on Drawings

The following article describes a way to use neural network models in mechanical engineering. Developing such methods is currently one of the most urgent tasks for specialists in this field, as it opens up prospects for automatically performing some of the functions that are currently performed by people. The paper describes a method that allows automated finding and recognition of the maximum diameter of a rotation part on drawings. In the course of the work, a new model was developed based on the YOLOv5 neural network model for recognition objects on images and videos, for recognition diametrical dimensions on drawings, and then a computer program was written in the Python programming language. The main purpose of the program is to use the developed model and the EasyOCR optical character recognition library to determine the maximum diameter value on drawings and enter the resulting data into an Excel spreadsheet. In the course of further analysis of the obtained characteristics and testing, both the developed model and the program showed their high efficiency. The results obtained have good prospects for their further application, for example, as a part of complex programs for automated equipment selection, determining the rate of material consumption for parts and grouping.