Variable-baseline disparity matching for 3D measurement and buried pipeline depth estimation

This study proposes a method to measure the ground surface in three dimensions using disparity matching of aerial images from disaster-affected areas. It also estimates the depth of shallow buried pipelines after an earthquake based on the 3D information obtained. Existing methods for measuring underground pipelines have several problems, such as high costs, long processing times, restrictions on pipeline material types, and limited measurable areas. The proposed method uses only two aerial images to generate a 3D map of the ground surface through disparity matching, enabling fast and wide-area measurements in disaster-affected regions. To verify the effectiveness of the proposed method, a small-scale observation environment has been developed. This system can simulate arbitrary seismic events and measure the response of underground pipelines. This innovation enables efficient acquisition of data before and after earthquakes, which has been difficult to obtain until now. In accuracy evaluation experiments, the system achieved displacement measurements with an error of less than 2 mm before and after vibrations, as well as post-vibration ground height measurements with an error of less than 1 cm. Furthermore, for the estimation of the depth of the buried pipelines after vibration, we explored regression analysis methods to estimate the depth of underground pipes based on the three-dimensional shape of the ground. As a result, using Random Forest Regression, we achieved a prediction error of less than 5 mm in the validation data set.