Laser scanning systems are employed in a wide variety of applications, such as digital building model generation, industrial site modeling, cultural heritage documentation, and other civilian and military needs. Depending on the mounting platform used, laser scanners can be classified into two groups: static or kinematic. A static laser scanner usually refers to a terrestrial laser scanner that is mounted on a tripod, while a kinematic laser scanner could be mounted on a fixed-wing plane, helicopter, or ground/terrestrial vehicle. A complete 3D model for a given site typically cannot be derived from a single scan. Therefore, several scans with significant overlap are needed to cover the entire site and also to obtain better information about the site than is possible from a single scan. When a static laser scanner is utilized to collect several scans to achieve complete coverage of a site of interest, the collected scans are referenced to different local frames that are associated with the individual scanner locations. Hence, a registration process must be performed in order to combine the several laser scans. The main goal of the registration process is to estimate the transformation parameters which determine the geometric variations between the reference frames of the collected datasets from different locations. The scale, shifts, and rotation parameters are generally used to describe such variations. The main objective of this thesis is a registration procedure for two overlapping static terrestrial laser scans using linear features that have been automatically extracted and matched in the available scans. The invariant characteristics of linear features in 3D space are utilized for establishing hypothesized matches of linear features between the overlapping scans; hence, the registration is performed in a pairwise scenario. Three alternative matching strategies are proposed to form correspondences between the linear features in overlapping laser scans. Another objective of this research is a segmentation/extraction procedure for low-level geometric features, such as planes,3D lines, and cylinders, from static and mobile laser scanning data, which is conducted through a region-growing segmentation process. The main characteristics of the proposed segmentation procedure are utilization of the proper parameterization models for the features of interest as well as consideration of the noise level and the point density variation within the laser scanning data. Experiments were performed to verify the outcome of the proposed segmentation and registration procedures. The segmentation experiments were conducted on static-terrestrial, mobile-terrestrial, and airborne laser scans, and an existing quality control procedure was used to evaluate the segmentation results. In the registration experiments, the proposed matching strategies were tested with real datasets of static laser scans captured over sites that are rich with linear features.