A Target Detection Algorithm for Marks Layout in Detail Drawings of Steel Structure
摘要
Detail drawings play an essential role in the steel structure manufacture as the basis for guiding manufacture and installation. Detail drawings are generally generated through software automatically nowadays. However, the current drawings created automatically by software are not of high quality, and the marks are overlapped, unattractive and unclear in the automatically generated drawings, which require considerable effort to manually adjust the position of the marks. In this paper, we developed an artificial intelligence algorithm for marks layout based on target detection algorithm. First, the marks layout problem is defined as the target detection problem. Then the Faster RCNN algorithm is refined on the basis of the unique characteristics of the marks layout problem, so that the algorithm is greatly strengthened in terms of accuracy and efficiency. It has been verified that after learning from the marks library, the marks generated by the proposed algorithm meet the requirements of manufacture. The aesthetics and clarity are significantly improved, and the workload of manual marks adjustment is dramatically reduced.