The term dataset applies to a collection of georeferenced raw input that are the primary source of data to run processing based on images or point clouds.
A dataset can therefore name one of the following in delair.ai :
• A collection of contiguous pictures, acquired with enough overlap (usually between 70% and 80%), in order to run photogrammetry processing for 2D orthomaps or 3D models to be generated over a continuous area of interest. These pictures can be of RGB type, or a collection of single multispectral images (R,G,B,RE,NIR), or a set LWIR thermal images. These pictures can carry metadata. (current categories supported in delair.ai)
• A collection of georeferenced points resulting from the processing of a LiDAR or laser scanner sensor raw data, defining a tridimensional point cloud. This point cloud can be optionally accompanied by a RGB dataset notably for the purpose of colorization of the point cloud. In such case, the dataset for this project is the combination of both LiDAR and RGB data over the same geographical area or interest. A dataset can be accompanied by extra data for georeferencing such as PPK files or GCP positions.
• A single image, orthomosaic, point clouds are datasets in delair.ai.