Filtering Your Data


We recognized a lot of confusion regarding filters and how to use them. In an effort to make choosing the right filter for your project easier, Propeller recently re-classified how we describe the filters available when uploading and processing data. It’s important to note the filters have not changed and still function the same way. 
Filtering is an automated algorithm and is not perfect. We recommend using either the No Filter or Equipment filter and combining this with the Terrain Cleanups tool to ensure you only remove items that would affect measurements onsite.That said, if you have been using an Advanced filter, have tried it, and are satisfied with the results, please feel free to continue using it.

How does it work?

A point cloud is a set of points in a 3D system. Each point represents a location in space and represents objects and terrain. When processing drone imagery, each pixel within the image has associated coordinates (X, Y, and Z). After your images are stitched together, these pixels are reconstructed into 3D point clouds of varying densities to replicate the area you just flew over. Filtering groups points within the constructed point cloud model into a specific class based on their characteristics. By bundling them together, we can remove them from the final elevation model.

For a point to be classed as something other than “ground,” it has to meet all three of these criteria:

  1. Point Distance (m): Sets limitations on how far a point can be from the ground. The greater the distance, the more likely the point belongs to a raised obstruction.
  2. Hold Size (m): Defines an area in which groups of points should be left untouched. This value can vary from a stockpile to a large building.
  3. Point Angle (degrees): Defines a line between the ground and a raised feature. The greater the angle, the greater the likelihood that the point belongs to a vertical feature. 

How do you choose the filter best for your survey?

Filtering obstructions on site is an important aspect of monitoring assets and job progress. Whether it’s a shrub impeding a calculation of bare earth or a vehicle parked on a surface you want a gradient read on, filters help minimize such obstructions. Think about the current obstructions onsite you are most concerned about affecting your data. The list below gives you an idea of what we can filter from your 3D surveys. When choosing a filter, remember that bare earth can only be rendered where the camera can see it. If the camera cannot see it, the filtering will cut away as much of the obstruction as possible but may not provide a completely flat surface below. 

 

No filter

This option is best if you want to keep all data in your survey unaffected by filtering. Choosing this option does not apply a filter.

 

Equipment

The best filter to choose if there are stockpiles you are measuring onsite. It removes anything up to and including the size of vehicles, haul trucks, or conveyors. It is unlikely to affect stockpile volumes.

 

Equipment and vegetation

Removes everything the Equipment filter removes, plus everything up to and including large trees. This can also include the tops of stockpiles. You’ll see bare earth underneath trees, shrubs, etc. The exception is when the vegetation is dense enough that the camera cannot see the ground over a large area. 

 

Equipment, vegetation, and structures

Removes everything the above filters remove, plus everything up to and including the size of large structures and buildings. 

 

Everything

Removes stockpiles entirely and everything else to produce a “bare earth” model. It cannot filter large areas of dense vegetation where the ground is not visible to the camera. Only use this filter on flat sites. 


I still can't do it!

We wrote these articles to equip you with everything you need to get the job done on your own, but we understand that sometimes this isn't sufficient.

If you're stuck, you can connect with our support team by clicking the support button on the top right corner of your user portal.