No abstract

Abstract-A lot of research has been done in the area of Simultaneous Localization and Mapping (SLAM). It is particularly challenging in underwater environments where many of the self-localization methods used at land no longer work. This paper presents a method to use a rotating scanning sonar as the sole sensor to perform SLAM in harbors or natural waterways. We present a feature extraction process which is capable of extracting walls of arbitrary shape. These extracted wall features are then used to perform SLAM using the wellknown FastSLAM algorithm. We show that SLAM is possible given this type of sensor and using our feature extraction process. The algorithm was validated on an open water test site and will be shown to provide localization accuracy generally within the error of the GPS ground truth.

T-splines have recently been introduced to represent objects of arbitrary shapes using a smaller number of control points than the conventional non-uniform rational B-splines (NURBS) or B-spline representatizons in computer-aided design, computer graphics and reverse engineering. They are flexible in representing complex surface shapes and economic in terms of parameters as they enable local refinement. This property is a great advantage when dense, scattered and noisy point clouds are approximated using least squares fitting, such as those from a terrestrial laser scanner (TLS). Unfortunately, when it comes to assessing the goodness of fit of the surface approximation with a real dataset, only a noisy point cloud can be approximated: (i) a low root mean squared error (RMSE) can be linked with an overfitting, i.e., a fitting of the noise, and should be correspondingly avoided, and (ii) a high RMSE is synonymous with a lack of details. To address the challenge of judging the approximation, the reference surface should be entirely known: this can be solved by printing a mathematically defined T-splines reference surface in three dimensions (3D) and modeling the artefacts induced by the 3D printing. Once scanned under different configurations, it is possible to assess the goodness of fit of the approximation for a noisy and potentially gappy point cloud and compare it with the traditional but less flexible NURBS. The advantages of T-splines local refinement open the door for further applications within a geodetic context such as rigorous statistical testing of deformation. Two different scans from a slightly deformed object were approximated; we found that more than 40% of the computational time could be saved without affecting the goodness of fit of the surface approximation by using the same mesh for the two epochs.

No abstract

No abstract