You have a point cloud with millions of data points, but you need discrete topographic points for your project. The traditional option? Sit down and place points manually. One by one. Sparks offers another way.
Key Takeaways
- Sparks automatically generates topographic points from LiDAR data in minutes
- Two modes: Interpolated (smooth, regular grid) and Real LiDAR (exact measured elevations)
- Obstacle detection identifies and corrects points on trees/buildings using MAD analysis
- Direct TIN surface and contour line generation from created points
- Classification filtering lets you use only ground-classified points
The Job Nobody Wants to Do
Anyone who has worked in surveying knows this process: you have a point cloud with millions of data points, but you need to convert it into discrete topographic points for your project. Points that represent the terrain cleanly. Points you can use to generate contour lines or TIN surfaces.
The traditional option is to sit down and place points manually. One by one. Click, verify elevation, adjust if necessary, next point. On a medium-sized survey, we're talking about hours of repetitive work.
Viizor Desktop includes a tool called Sparks that tries to simplify this process. It's not perfect, but it can help quite a bit.
What Does Sparks Actually Do?
Sparks analyzes a section of the point cloud that the user defines and places topographic points automatically. The process works like this:
- You define an area of interest (AOI) — You draw a polygon over the zone you want to process
- You configure the parameters — You choose the grid spacing and point type
- Sparks analyzes the terrain — It examines the LiDAR points inside the polygon
- Points are generated — Topographic points appear ready to use
The algorithm uses Delaunay triangulation to understand the terrain geometry. This means it doesn't simply place points on a rigid grid ignoring what's below — it tries to adapt to the actual surface.
Two Working Modes
Interpolated Mode
Regular grid, smooth result
Creates a regular grid and calculates elevation by interpolating from nearby LiDAR data. Results in uniform point distribution. Useful for consistent density or contour generation.
Real LiDAR Mode
Irregular, exact elevations
Looks for actual LiDAR points near each grid position. Uses exact X, Y, Z coordinates from points that actually exist. Distribution is less regular, but elevations are measured data.
Each mode has its use depending on what you're doing. Interpolated mode works well when you need smooth contour lines. Real LiDAR mode preserves the actual measured elevations when that precision matters.
Obstacle Detection
One thing Sparks tries to do is identify points that probably aren't terrain. If a point has an elevation very different from its neighbors — as would happen if it falls on a tree or building — the algorithm detects it and tries to correct it.
How It Works
The process uses Median Absolute Deviation (MAD) to identify outliers. If a point deviates more than 3 × MAD from its neighbors (with a minimum threshold of 0.5 meters), it's considered suspicious.
When Sparks finds one of these points, it looks for the lowest point within a 15-centimeter radius. If it finds one, it uses that elevation instead. The idea is that if there's a tree, some points probably passed between the leaves and reached the ground.
If it finds nothing in that radius, it simply discards the point.
It's not a foolproof system. A very dense forest or a structure that completely covers the ground will still be problematic. But in many situations, it helps obtain a cleaner result without having to review each point individually.
Grid Spacing
You can choose between two options:
Auto: The algorithm calculates spacing using the formula √(0.3% of polygon area). In an area of 10,000 square meters, this gives approximately 5.5 meters of spacing. Additionally, it applies a small random variation (±5%) so the points don't end up perfectly aligned — something more similar to what a surveyor would do working manually.
Custom: You define the exact spacing. If you need points every 2 meters because your project requires it, that's what you'll get.
Note: The system warns you if you're going to generate more than 5,000 points. Creating many points consumes memory and can slow down the viewer. It's not an absolute technical limit, but a warning so you can make an informed decision.
Polygon Alignment
A detail that might go unnoticed: the point grid aligns with the longest side of the polygon you drew. This makes the points follow the natural orientation of your work area instead of ending up in an arbitrary north-south grid.
Additionally, points are placed inside a slightly contracted polygon (5% inward). This prevents points from ending up exactly on the edges, where data tends to be less reliable.
Classification Filtering
If your point cloud has LAS classification, Sparks lets you filter which points it uses for analysis. You can choose, for example, to use only points classified as "Ground" (Class 2). This is useful when you've already done automatic or manual terrain classification and want to make sure the topographic points represent only the ground.
The available classes depend on what exists in your file — Sparks reads the data and shows you the actual options.
TIN Surface Generation
The points Sparks generates don't stay loose. You can select a group of points and create a TIN surface (Triangulated Irregular Network) directly. The system uses Delaunay triangulation to connect the points and creates a mesh representing the terrain.
This TIN surface has some useful features:
- Edge deletion: If there are very elongated triangles at the edges (common when point distribution is irregular), you can delete individual edges with a click
- Automatic regeneration: If you move or delete one of the original points, the TIN updates
- Edit history: There's an undo system (Ctrl+Z) for TIN modifications
Generated TINs can be exported to standard formats for use in other civil design programs. See From Point Cloud to Civil 3D for a complete walkthrough of the export options.
Contour Line Generation
From a TIN surface, you can generate contour lines. The system lets you configure:
- Interval between contours (for example, every 1 meter)
- Interval for index contours (for example, every 5 meters)
- Whether you want to show labels with elevations
- Line smoothing level
- Colors for normal and index contours
The contours are calculated by intersecting horizontal planes with the TIN surface. It's basic geometry, but the result is useful for visualization and for exporting to other formats.
What Works Well and Where It Has Limitations
What Works Well
Sparks is particularly useful on relatively open and uniform terrain. Clearings, fields, open areas, construction surfaces. In these cases, it can save you a significant amount of manual work.
It also works reasonably well on terrain with scattered vegetation, where the outlier detection algorithm can identify and correct points that fall on trees or bushes.
Where It Has Limitations
In areas with very dense vegetation, the result will require more manual review. The algorithm looks for low points to find the ground, but if there's no LiDAR penetration to the terrain, there's not much it can do.
On terrain with abrupt elevation changes — like steps, retaining walls, or excavation edges — interpolation may smooth out details you wanted to preserve. "Real LiDAR" mode helps, but doesn't guarantee capturing all discontinuities.
And like any automatic tool, it doesn't replace professional judgment. The generated points are a starting point, not necessarily the final product.
The Time Investment
Using Sparks on a typical project takes a few minutes. Draw the polygon, adjust parameters, wait for generation. Compared to placing hundreds or thousands of points manually, the difference is considerable.
What you do with the time you save is your decision. You can use it to more carefully review problematic areas. You can process more projects. You can go home earlier.
To Wrap Up
Sparks isn't magic. It's an algorithm that applies some known geometric techniques to automate a repetitive task. It works well in many cases, has limitations in others, and always requires a professional to review the result.
But if you've ever spent an entire afternoon placing topographic points one by one over a point cloud, you'll probably see the value in a tool that does that work in minutes.
Sparks comes included in Viizor Desktop as part of the terrain analysis toolset. It's not an additional module nor does it require a special license — it's available to all users.
Turn Hours Into Minutes
Sparks generates topographic points automatically. Included with Viizor Desktop — no extra cost.
$540 One-time payment
Windows 10/11 • No credit card required for trial