Example Request

Currently , the LidarTopDown (LTD) pipelines accept two types of data inputs: LiDAR and Aerial. To use both LiDAR and Aerial, please reach out to a Scale representative for further details.


The input is a series of Frame objects. Scale will take LiDAR point clouds as input, and render a top-down (bird’s eye view) tile image of the point clouds. These images are then used to create imageannotation subtasks, with lidar camera context attachments.

To submit data of this type, upload lidar Frame objects under attachments, as shown below:

 attachments: [


For aerial imagery, the input is a bird’s eye view image and optionally camera images for context. The main image is taken from the attachment property, and the context images will be taken from context_attachments. To submit data of this type, provide the URL of the image file under attachment:


Additionally, provide each non-primary image that can be referenced during labeling as a context attachment.

Note, in the case of a LiDAR task, this information is processed from the Frame objects provided under attachments. In the case of an Aerial task, this information must be submitted under the context_attachments property.