filters.statisticaloutlier.rst

filters.statisticaloutlier

The Statistical Outlier filter passes data through the Point Cloud Library (PCL) StatisticalOutlierRemoval algorithm.

filters.statisticaloutlier uses point neighborhood statistics to filter outlier data. The algorithm iterates through the entire input twice. The first iteration is used to calculate the average of the distances from each point i to its k nearest neighbors or AD_i. The second iteration is used to identify outliers based on the distribution of AD_i values. Points with an average neighbor distance greater than the mean plus 2 standard deviations are classified as outliers. The value of k can be set using $\tt mean\_k$. By default, the distance threshold is set to $\overline{AD}_{i} + 2 \hat{\sigma}$, but a value other than 2 can be chosen using $\tt multiplier$.

Before outlier removal, noise points can be found both above and below the scene.

After outlier removal, the noise points are removed.

See [Rusu2008] for more information.

Example

{
  "pipeline":[
    "input.las",
    {
      "type":"filters.statisticaloutlier",
      "mean_k":"12",
      "multiplier":"2.2"
    },
    {
      "type":"writers.las",
      "filename":"output.las"
    }
  ]
}

Options

mean_k

Mean number of neighbors. [Default: 8]

multiplier

Standard deviation threshold. [Default: 2.0]

classify

Apply classification labels? [Default: true]

extract

Extract ground returns? [Default: false]