Automated code compliance checking is an effective approach for assessing the quality of building information modeling (BIM) models. Various automated code compliance checking systems have emerged, wherein users need to input all information accurately according to BIM modeling guidelines, in order to ensure the accuracy of checking results. However, as this process involves human inputs, it is difficult to ensure that each input is accurate. In the case of errors or missing inputs, the checking results will be erroneous. Although automated checking systems can be developed accurately, it is difficult to apply these systems practically. Therefore, this paper proposes the application of AI technology to recognize BIM objects and visible properties, in order to improve the operability of automated code compliance checking. The two necessary elements – object names and properties – could be automatically extracted to a certain extent, following the application of the proposed method to the automated code checking process. The error rate of the input could also be reduced, thus making the application of the code checking system more practically feasible. The proposed recognition method for BIM objects and visible properties is also expected to be used widely in BIM-based building e-submission systems and BIM-based forward designs.
Read full abstract