CloudCutter_E57: The First Practical Breakthrough

From a workflow idea to a working process

The earlier problem was simple to state: engineering teams often need a small area from a very large registered point cloud, but the available workflow encourages them to load or duplicate far more data than necessary.

The first CloudCutter_E57 breakthrough was a practical pipeline that separated preview from export.

The core principle

The application reads the source sequentially and retains only a controlled sample for visualization. The user works with that sampled representation to define spatial selections. The final export then returns to the original file and writes only the points that satisfy those selections.

This architecture provides two important properties:

• preview memory depends on the sampling policy, not directly on the total number of source points;
• exported files can retain full-resolution source points inside the selected regions.

What the first implementation demonstrated

• Direct work with E57 source data.
• A lightweight preview suitable for spatial selection.
• Selection of areas of interest without constructing the entire cloud in RAM.
• Streaming extraction of selected points.
• Preservation of coordinates and available point attributes requested for export.
• Production of smaller E57 files for downstream engineering work.

What CloudCutter was not intended to replace

CloudCutter was not created as a scanner-registration system or as a substitute for professional suites such as FARO Scene, Leica Cyclone, Autodesk ReCap, or AVEVA Point Cloud Manager. Those products solve broader tasks.

CloudCutter addressed a narrower gap: the preparation of large registered datasets for local engineering use.

Why the breakthrough mattered

The technical achievement was not simply opening an E57 file. It was establishing a scalable workflow in which data volume did not force the entire source into memory. That principle later became the basis for multi-file processing, additional formats, 3D preview, repeated export, and automated multi-output generation.

The following publication shows the first version in operation.