Outlier removal - PIX4Dmatic

The statistical outlier removal filter detects points whose distance to their closest neighboring points is deviating from the typical distance over the whole point cloud.

The filter checks each point in your data to see how close it is to the points around it. It then compares this distance to the average distance of all points to their neighbors. If a point is much farther away from its neighbors than the average, it's considered an outlier and is removed.

The filter is controlled by two main parameters: distance factor and number of neighbors.

• Number of Neighbors (From 1 to 20. By default 8): This setting defines how many nearby points the filter considers when calculating the average distance for each individual point.

  • A high setting for the number of neighbors makes the filter more robust to local variations in point density because it considers a larger group of points.

  • A low setting is useful for projects with highly varying point densities or when you need to preserve fine details, as it only considers a small, localized number of points.

• Distance Factor: (From 0 to 10. By default 1.5): This is the threshold for flagging a point as an outlier. It determines how far away a point must be from its neighbors relative to the average distance of the entire point cloud.

  • A high setting means the filter is very forgiving; a point must be extremely far away to be considered an outlier. This will remove fewer points.

  • A low setting makes the filter stricter; a point does not have to be very far away to be removed. This will result in more points being filtered out.