Centerline Extraction

Function Description: Extracts the tunnel centerline as the baseline for point cloud unrolling. Centerline Extraction must be completed before running Unrolling.

Parameters

  • Source Class: Tunnel point cloud class.

  • Method: Supports three extraction modes: Segmented Matching, Geometric Center, and Circle Fitting.

  • Step Size: Mileage interval between extraction frames.

  • Thickness: Cross-section point cloud thickness. This value can be appropriately increased when point clouds are sparse.

  • Error Threshold: When the extraction error of the current frame exceeds the threshold, discard the extraction result of that frame.

  • Smoothing: When checked, apply smoothing to the extraction results.

Segmented Matching

Suitable for complex tunnel contours (non-single circular arc composed). By matching the tunnel cross-section contour point cloud, take the geometric center of the projection segment as the centerline position.

Note: When Unrolling uses the Segmented Projection method, Segmented Matching must be selected; otherwise, the projection results may deviate from expectations.

Geometric Center

A universal centerline extraction method that can obtain ideal centerline results without considering the specific contour shape. However, the centerline extracted by this mode is not suitable for the Unrolling function.

Circle Fitting

Suitable for tunnel contours with a single circular arc shape. Perform least squares fitting on the point cloud within the specified azimuth range, and take the circle center as the centerline position.

  • Azimuth: Calculated from the zenith direction on the projection cylindrical surface, with clockwise direction as positive.

Note: At tunnel entrances and exits, the center point position may shift due to vegetation point interference. It is recommended to run Classification by Deep Learning first.

results matching ""

    No results matching ""