Reacting to clicks in point cloud models

While point clouds doesn't have a concept of nodes, applications can detect clicked points. Just like with CAD models, Cognite3DViewer.getIntersectionFromPixel is used to determine world position and point from a mouse click.

Detecting clicked positions

The following example detects intersections at clicked positions and marks the intersected positions using a red sphere.

Live Editor

Getting information about object assigned to a clicked point

Using Cognite's Annotations API or Cognite's 3D Contextualization Editor in CDF , a user can define objects within a point cloud model. In order to retrieve information about a clicked object, users may get the annotationId or the volumeInstanceRef or the assetRef that is assigned to each point.

The annotationId and assetRef or volumeInstanceRef and assetRef are available in the intersection object returned from viewer.getIntersectionFromPixel.

An annotation id of 0 and a volumeInstanceRef of undefined means that the clicked point is not assigned to any object. assetRef will only be defined if the annotation associated with the point is linked to an asset in CDF. The objects in this point cloud example have no assets associated to them.

Live Editor

Cognite core data model point cloud volume click example

viewer.on('click', async event => {
  const intersection = await viewer.getIntersectionFromPixel(
    event.offsetX, event.offsetY
  );
  if (intersection) {
     alert(`Clicked object with point cloud volume reference externalId ${intersection.volumeInstanceRef.externalId} and assetId ${intersection.assetRef?.id}`);
  };
});