Filtering Operations

There are many different combinations of filtering operations. In this section, Filter by threshold and Filter by label are used as examples and can be applied flexibly according to actual needs. The possible filtering operations include, but are not limited to point clouds, poses and labels. This section takes point clouds as an example.

../../../_images/Example_combinations.png

Figure 1 Example of procedure for filtering operations

  1. The procedure Filter by Threshold

  2. The procedure Filter by Label

Process Description

Filtering by label

  1. Validate Labels and Output Flags step and Label File generate an array of flags with the same number of tags as the input, with the values True and False.

  2. The Filter step keeps the input with the flag True and filters out the input with the flag False ; if Reverse Bool List is checked, the result is reversed.

Filtering by threshold

  1. The Stratify Values by Thresholds step sorts the input array of values according to a threshold value. Less than the threshold is the previous category label and greater than or equal to the threshold is the next category label.

  2. Use the Filtering by label step combination to filter the labels and achieve the threshold filtering effect.

Example of How to Achieve The Effect

Input point cloud

../../../_images/Origin_cloud.png

Figure 2 Input Point Cloud Image

Process flags

../../../_images/Thresh_value.png

Figure 3 Input of Stratify Values By Thresholds

Here the threshold is 2000 and the labels are ng, ok. The required label is ok .

../../../_images/Thresh_result.png

Figure 4 Input of Validate Labels and Output Flags

Retain the point cloud class with the flag True .

../../../_images/Label_result.png

Figure 5 Input of Filter

Resulting point cloud

../../../_images/Result_cloud.png

Figure 6 Resulting Point Cloud Image