Data Manipulation: Light Detection and Ranging – Essay Example

The paper “Data Manipulation: Light Detection and Ranging” is an excellent example of an essay on information technology.
The process of manipulating data using a variety of techniques, including Light Detection and Ranging (LIDAR), is known as surface modeling and will be used to visualize the front of Kingston University (Penrhyn Road Campus) and analyze using different measurements. To accurately model a surface, scans must be performed from different positions. These scans are then registered through alignment into a coordinate system by comparing cloud appears, objects or geometric features between different scans using the functionality of cloud constraints or targets.
As the scans taken in 2012 were poor quality due to bad weather conditions, data was used from the previous year. The data included four scans, which were cleaned, registered and modeled with Leica Cyclone. To register the scans, cloud mesh between each set of scans (6) and targets (8) with lower errors values (up to 0.005m error) was analyzed. The registration process began by adding the shared targets between scans using the Auto-add Constraints function, followed by optimizing the alignment, resulting in adding the targets of the four scans. Next, the mesh was added manually by picking points between two scans and using cloud constraint function to automatically add mesh using current registration. To further improve the results, the subsampling percentage of cloud constraint changed from 3% (default) to 100%, and the optimizing alignment was repeated, resulting in reduced error values. A number of targets were deleted due to their high error value, resulting in highly accurate registration. The final step was freezing and creating a scan world using registered scans to begin modeling and analyzing.
Modeling generally involves fitting geometric objects to point clouds, depending on the topology of those points. This process is time-consuming as great detail is required and skills of sub-selection create and editing object functions are necessary, in addition to artistic ability to produce highly standardized visualization models. Before beginning modeling, the front fence must be divided into three parts to avoid any errors caused by focusing on a single side of the building. Fit fence function was used to model walls and windows, while right side fence cylinder objects were used to represent objects. After modeling the entire building, it was necessary to merge the three parts to produce the final model of the building. However, difficulties emerged during the modeling and merging process, particularly in aligning the walls, as each wall was created using a different operation of region growing function. In addition, the windows of the second floor of the entrance in the oblique angle were problematic. The extrude function was used to add final touches to the building, giving it a more aesthetic feeling.
Although alignment is much more convenient with other modeling software, such as Google Sketch Up, other options often suffer from poor location accuracy. To achieve low error values in the registration process, scan positions should be relatively close and share 25% (scans 1-2, 2-3, 3-4). When software is only available on university campuses, users have difficulty working in their own time. The greater the amount of time allocated to the modeling process, the greater detail will be possible, producing higher quality than models created in short timespans.