It is often thought that a survey project ends when the vessel leaves the site. But in fact, this is just a start of the data’s journey. The raw data gathered from the field still needs a lot of attention to be utilizable in asset management.
Post-processing finalizes the accuracy of the data
VRT’s strict quality requirements demand a lot of knowledge from the post processing crew. Because the data is used in asset maintenance, the good accuracy is essential, and it is achieved with the combination of conscientious field work and post processing.
Our survey vessels are equipped with GNSS (Global Navigation Satellite System) antennas and inertial navigation system, which record vessel trajectories (information about the vessel’s positioning) during the field work. After the field work, the data is processed with PPK (Post-Processing Kinematics) calculations to reach few centimeters accuracy. The calibration of sonar and laser scanner include the calibration lines analyzing to adjust heading, pitch and roll angles. After that the point clouds are exported and analyzed yet again by manually comparing the datasets.
“Even after the first two calculation phases there can be some errors from GNSS dead zones or SVP (Sound Velocity Profile) flaws, which are detected in the final check-up. By adjusting the trajectory information or SVP graph the errors can be fixed, if you understand profoundly how the adjustments affect the point cloud”, Data Processing Specialist Lassi Aaltonen explains.
Cleaning of the point cloud adds visuality to the data
As all our datasets go through the same thorough post-processing, it is easy to attach the datasets together to make a complete model of the area or structure. There are no gaps between the datasets and the quay decks are lined beautifully together. Still the 3D data isn’t understandable before the outlier points are removed, which is called point cloud cleaning.
Point cloud cleaning can be automatized to some extent, but when we want to achieve the accuracy eligible to structural inspections, the cleaning by a trained professional is essential and worth the time spent. It ensures that the structures and possible damages, or even minor deviations, can be spotted, and that the 3D model is illustrative to also an untrained eye. Even the dataset with super high resolution does not tell much without the data specialist’s touch. For example, pile structures under a quay are just blurry if the data is not properly cleaned.
“We use multiple software to have the best result for every type of data and structure. The fine-tuning of the point cloud so that it can be used in structural inspections, and it is informative enough in asset management, requires experience. You can say that the well-cleaned and visualized 3D data is a work of art”, Project Engineer Eemeli Ahonen concludes.
If you want to see how VRT’s 3D Survey results differ from the average survey data, contact our specialists to hear more. We have experience in all kind of underwater infrastructure from energy utilities to ports and bridges.
Read also the other parts of the blog series 3D Survey as a Service: