cluster extraction in point cloud

4 views (last 30 days)
Pavan Kumar B N
Pavan Kumar B N on 31 May 2021
Answered: Vidip on 23 Jan 2024
Hell All,
How to remove unwanted clusters from point cloud data and fit bounding box fot object detection and tracking.
Thankyou,

Answers (1)

Vidip
Vidip on 23 Jan 2024
To remove unwanted clusters from point cloud data and fit bounding boxes for object detection and tracking, you typically need to follow a series of steps. Here's a general approach:
  1. Preprocessing: Start by filtering the raw point cloud data to remove noise and outliers. It helps in enhancing the quality of the point cloud for further processing.
  2. Segmentation: Segment the point cloud to separate the points of interest from the background. This can be achieved using techniques such as Euclidean Cluster Extraction to isolate distinct objects.
  3. Clustering: After segmentation, apply clustering algorithms such as DBSCAN or Euclidean Cluster Extraction to group the remaining points into clusters based on their proximity. Each cluster is a candidate for an object of interest.
  4. Removing Unwanted Clusters: Identify and remove unwanted clusters based on criteria such as size, density, or location. For example, you can discard clusters that are too small to be considered objects or those that are in regions not relevant to the analysis.
  5. Bounding Box Fitting: For each of the remaining clusters, fit a bounding box that encapsulates all the points within the cluster. The bounding box can be axis-aligned or oriented, depending on the requirements of the object detection and tracking task.
  6. Object Detection and Tracking: Use the fitted bounding boxes as inputs for object detection and tracking algorithms. The bounding boxes provide initial estimates of the object locations and extents, which can be refined and tracked over time.
For further information, refer to the documentation links below:

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!