LiDAR Processing

Point Cloud Reconstruction

Raw LiDAR data includes beam number, beam azimuth, range measurements (relative to the LiDAR unit) and intensity/reflectivity. For mobile LiDAR (i.e., LiDAR mounted on a moving platform), mounting parameters from system calibration and vehicle trajectory are required to reconstruct a point cloud in a reference frame (e.g., Earth-centered, Earth-fixed coordinates). Figure 1 illustrates the process.

Combing the LiDAR range measurement and orientation of the laser beam establishes the coordinates of a point, $I$, captured at time, $t$, relative to the laser unit frame, $r_I^{lu(t)}$. The GNSS/INS integration provides the trajectory information, i.e., position, $r_{b(t)}^m$, and orientation, $R_{b(t)}^m$, of the GNSS/INS body frame relative to the mapping frame. The system calibration determines the lever arm, $r_{lu}^b$, and boresight matrix, $R_{lu}^b$, relating the laser unit frame to the GNSS/INS body frame. The 3D coordinates of the point, $I$, in the mapping frame can then be derived based on the following equation.

$$r_I^m = r_{b(t)}^m + R_{b(t)}^m r_{lu}^b + R_{b(t)}^m R_{lu}^b r_I^{lu(t)}$$

Figure 1. LiDAR point cloud reconstruction.

According to the process described above, the quality of LiDAR point cloud is contingent on laser range measurement accuracy and system calibration and trajectory quality.