The Metro Construction Drawings Recognition For Risk Identification

Abstract

The most important task of risk identification in the metro construction safety management is to capture technical parameters by reading construction drawings manually. It is a heavy work, causing misreading, omitting information, etc. This paper introduces a novel method for automatic recognition of the real-life electronic metro construction drawings. It defines sequential, orientational, integrated recognition strategies by analyzing and summarizing drawings features, descripts recognition process by BNF, and designs four recognition algorithms to recognize typical graphic elements. Developing construction drawings recognition software and applying it in the metro project risk management in cities of Wuhan, Zhengzhou and Shenyang. It improve 10-20 times reading drawings efficiency.