A self organizing map optimization based image recognition and processing model for bridge crack inspection

The current deterioration inspection method for bridges heavily depends on human recognition, which is time consuming and subjective. This research adopts Self Organizing Map Optimization (SOMO) integrated with image processing techniques to develop a crack recognition model for bridge inspection. Bridge crack data from 216 images was collected from the database of the Taiwan Bridge Management System (TBMS), which provides detailed information on the condition of bridges. This study selected 40 out of 216 images to be used as training and testing datasets. A case study on the developed model implementation is also conducted in the severely damage Hsichou Bridge in Taiwan. The recognition results achieved high accuracy rates of 89% for crack recognition and 91% for non-crack recognition. This model demonstrates the feasibility of accurate computerized recognition for crack inspection in bridge management.