On the Sensitivity of the Parameters of the Intensity-Based Stochastic Model for Terrestrial Laser Scanner. Case...

Area of Science:

• Geomatics Engineering
• Photogrammetry and Remote Sensing
• Geodetic Science

Background:

• Accurate least-squares (LS) solutions require precise stochastic properties of measurements.
• Terrestrial Laser Scanner (TLS) range variance is often modeled using intensity-dependent power laws.
• Current methods for determining these parameters are complex and may limit the use of empirical intensity models (IM).

Purpose of the Study:

• To investigate the impact of intensity model parameter variations on LS adjustment results using simulations.
• To assess the suitability of point-wise weighting in LS adjustments with derived variance-covariance matrices (VCM).
• To propose a simplified procedure for determining range variance for VCM computation.

Main Methods:

• Closed-loop simulations with controlled geometry and stochastic models.
• Analysis of least-squares adjustments for determining control points (CP) from B-splines curves.
• Investigation of constant variance assignment for homogeneous point clouds.

Main Results:

• Assigning a constant variance to points with homogeneous properties does not negatively impact the a posteriori variance factor or LS solution efficiency.
• Simulation findings remain valid for complex geometries in real-world scenarios.
• A simplified procedure for range variance determination is proposed, streamlining VCM computation.

Conclusions:

• A constant variance approach is a viable simplification for processing homogeneous TLS data.
• The proposed method simplifies VCM computation, making LS adjustments more accessible.
• This research offers a more efficient pathway for utilizing TLS data in geodetic applications.