Design of CAD Table Recognition Algorithm Based on Deep Learning
With the rapid development of the engineering and design industries,Computer-Aided Design(CAD)play an indispensable role in producing design drawings.However,traditional CAD systems have limitations in managing and extracting tabular data,especially when dealing with large-scale engineering drawings.To address this issue,this paper introduces a new automated method for extracting large table data from CAD drawings.By converting original CAD files into image formats and applying advanced image processing techniques along with Deep Learning model(SAHI algorithm and Cycle-CenterNet model),this method can effectively improve the ac-curacy and efficiency of table data recognition and processing.Experimental results show that,compared to direct extraction of table data,using this method significantly enhances the precision,recall,and F1 score of data extrac-tion.Future work will focus on optimizing the model architecture and enhancing its applicability and performance across various types of drawings.
万方数据
维普
