Image Recognition Empowers Intelligent Analysis of Industrial Drawings

This paper develops a specialized drawing data management software to address the challenges of information management and the application of drawing data in industrial production. This software integrates image segmentation algorithms based on OpenCV and text recognition algorithms based on CnOCR. It is designed to process two types of drawing data: part drawings and assembly drawings. First, the PDF files are converted into images. For part drawings, edge detection algorithms in OpenCV are utilized for segmentation, and recognition algorithms are employed to accurately locate and remove numbers, ultimately generating quality inspection diagrams. A text recognition algorithm is utilized for assembly drawings to extract data from the drawings, followed by feature processing and analysis to extract information regarding all components depicted in the drawings. The software also incorporates a feature for searching by component shape to facilitate the retrieval of drawing data. The findings suggest that the algorithm effectively recognizes drawing data, automates data statistics and quality inspection processes, elevates intelligent drawing data management, and holds substantial practical utility.