8.1 Understanding the LixelStudio Pipeline

What LixelStudio Does

LixelStudio is the point cloud processing software for the L2 Pro and K2. It takes raw scan data from the device and converts it into a usable, navigable point cloud through a sequence of automated stages. PortalCam data is not processed in LixelStudio. That device uses LCC Studio exclusively.

The output is a georeferenced, optionally colored, three-dimensional point cloud that can be exported into BIM, GIS, and survey workflows. There is no conversion path between LixelStudio output and LCC Studio output. If you need both a point cloud and a 3D Gaussian Splatting model from the same scan, both pipelines must be run separately from the original raw data. Do not run LCC Studio and LixelStudio processing jobs simultaneously on the same machine. Each application consumes significant GPU, RAM, and disk I/O. Process one pipeline to completion before starting the other, or use two separate workstations.

Pipeline Stages

Stage 1: SLAM Optimization

LixelStudio reconstructs the device trajectory and point map from the raw LiDAR and IMU data recorded during the scan. This is the computationally intensive phase that determines the overall accuracy and completeness of the result. The entire scan trajectory is held in RAM during this stage, which is why RAM capacity directly limits the maximum scan duration that can be processed.

SLAM optimization runs with no special mode by default (None mode), which provides the highest accuracy in stable environments with rich features. If processing fails in None mode, particularly with an "LIO trajectory drift" error, enable Robust Mode from Advanced Settings before concluding the data is unrecoverable. Robust Mode improves success rate in partially degraded or high-motion scenes but may reduce accuracy. Narrow Scene Mode is reserved for tunnels, mine shafts, and long corridors, and using it in a standard environment produces a processing failure.

Stage 2: Point Cloud Generation

Once the trajectory is established, LixelStudio generates the dense point cloud from the LiDAR range data along the reconstructed path. Point density settings configured before processing determine the resolution of the output. Higher density produces larger files and requires more processing time.

Stage 3: Coordinate Transformation

If RTK, PPK, or GCP data was collected during the scan, this stage applies that data to place the point cloud in a real-world coordinate system. Without georeferencing, the output exists in an arbitrary local coordinate frame. LixelStudio redesigned the coordinate transformation interface to support fast search of common and built-in coordinate systems, with clear and intuitive naming for easier identification.

Stage 4: Color Mapping

If camera data is available, LixelStudio maps color from the camera images onto the point cloud geometry. LixelStudio substantially improved the colorization algorithm, with more realistic color rendering, clearer feature and boundary definition, and significantly better performance in low-light environments such as basements, parking lots, and corridors.

The K2 produces real-time true-color point clouds from its 3 built-in cameras (2 fisheye + 1 forward, 200°×200° panoramic coverage) during scanning, with no separate coloring step required during processing. L2 Pro uses its 2 onboard 48 MP panoramic cameras (plus a 1 MP positioning camera) and color mapping is applied during the processing pipeline. Color mapping requires CUDA 11.6 or later and an NVIDIA GPU from the 20-series or newer. 10-series and older GPUs cause a coloring crash.

Workstation Requirements

LixelStudio lowered the entry hardware bar significantly compared to older LixelStudio releases. A workstation that previously could not run older LixelStudio releases will likely run Basic. Most professional workstations bought in the last four years now meet the Basic specification. AMD GPUs remain unsupported, and there is no driver update or workaround for them.

Basic is sufficient for K2 single-segment captures and small L2 Pro projects. K2 datasets are smaller and faster to process than L2 Pro data, so a Basic-spec workstation handles K2 production work comfortably for most use cases. Recommended is necessary for L2 Pro multi-session map fusion and any project with total scan duration above 60 minutes. SATA SSD increases processing time by 15 to 25%, so NVMe is the correct choice for any production workstation.

Processing Time

LixelStudio processing runs at approximately 20 to 30 times the scan duration on Recommended-tier hardware. This ratio applies to SLAM optimization on an NVMe SSD workstation. Color mapping adds additional time on top of the point cloud processing. Both the GPU and CPU are utilized during processing.

LixelStudio displays an estimated processing time in a popup window when you start a job. Use this estimate for scheduling. The actual duration varies with scan complexity, environment geometry, and workstation performance, but the estimate provides a reasonable planning baseline. K2 datasets typically process faster than L2 Pro datasets of comparable duration because the data is smaller and more uniform.

Software and Firmware Requirements

For new L2 Pro and K2 fleets running current firmware, use the current LixelStudio release. After any firmware update, verify LixelStudio compatibility before processing new data. LixelStudio is the current production release as of May 2026 and supports all current device firmware including L2 Pro Drone Mode (which previously required a current LixelStudio release).

Common Misconceptions

These are the misunderstandings that cause the most wasted time in the field and in processing. Each one is based on reasonable assumptions that happen to be wrong for this system.