The level of precision we achieve in any design or construction project is very dependent on our understanding of the data that represents current site conditions. Frequently, AEC teams use a combination of LiDAR/laser scanning, photogrammetry, and/or Global Navigation Satellite System (GNSS), chosen based on the project’s budget, desired level of accuracy, and area coverage constraints.
In particular, LiDAR and laser scanning are widely used in the engineering industry to create highly precise 3D documentation of infrastructure, buildings, and complex mechanical systems. The resulting point clouds can be integrated directly into BIM or CAD programs; however, their massive scale requires new approaches to data processing.
Point Clouds Break Traditional Data Processing Methods
- High-Density Scans. A 3D laser scanner can capture as many as two million data points per second. That means that in a single scan, surveyors can end up with millions if not billions of points. Large projects requiring hundreds or thousands of scans generate point clouds that are too dense for efficient transfer and storage, significantly increasing project costs.
- Computing Demands. Due to their sheer size, point clouds can consume significant computing. Handling point cloud data typically exceeds the capabilities of standard computers, requiring specialized high-performance hardware to avoid slowdowns and system failures.
- Muddy Data. Real-world point clouds often contain noise and outliers produced by equipment vibrations, reflective materials, and changing weather patterns. Left untouched, these outliers reduce the integrity of the point cloud; however, manually classifying points is a time-consuming process and produces inconsistent results.
Getting a Lift with Automation in the Cloud
Finding a solution to a problem is one step forward, but creating a solution that scales to different use cases is true success. Real-time processing requires distributed computing and cloud-based systems that can handle the demands, simplify processing, and enable easier interaction:
- Replace the hassle of multi-step data cleaning with an automated solution that requires minimal user input.
- Maintain local computing power by using cloud-based solutions that address storage challenges and minimize CPU strain, freeing system processing power for your team to focus on other critical tasks.
Cleaning Scanned Floor Data with RoboClean
When scanning very flat surfaces, such as concrete floors, RoboClean provides an easy-to-use, automated way to quickly remove noise from point clouds. With its primary focus on extracting floor slabs, RoboClean classifies both ground and non-ground (noise) points, delivering E57 cleaned files as well as LAS files that include all point classifications. To get started:
- Upload your noisy point cloud to BRYX.
- Run your data through RoboClean for automated cleaning. When configuring RoboClean, you can choose your desired cleaning intensity, specify the far or near distance from the scanner, and upload a DXF file that defines boundaries to be excluded from the cleaning operation.
- Access your results. The cleaned E57 files can be easily submitted to RoboFlat for concrete floor flatness/levelness analysis. With RoboClean and RoboFlat, scanning professionals have a streamlined solution for evaluating concrete floors for flatness and levelness.
In three steps, RoboClean generates clean, actionable files that can be used to guide your floor analysis. Start your free trial today by visiting BRYX.