Skip to content

Latest commit

 

History

History
204 lines (171 loc) · 14.1 KB

5_DataProducts.md

File metadata and controls

204 lines (171 loc) · 14.1 KB

UAS Data product types and processing levels

A scheme of data Levels is proposed to allow data users to clearly identify the types of data products within a UAS data package.

The following tables describe data levels, giving examples of data types and file formats. With ‘File name recommendations’, the *** represents a prefix for a flight or campaign. This prefix should represent the study site, acquisition date, and/or flight identification as appropriate.

Note that this list is intended to show examples for each data Level, and is not exhaustive of all possible products or file formats. csv file types should follow the ESS-DIVE Reporting Format for Comma-separated Values File Structure.

Level 0 (L0): raw data collected with the platform

Level 0 data are raw data directly from the instrumentation on board the UAS platform. In addition, this data should include basic any flight mission telemetry and trigger records that help users identify the location, time, data triggering events needed for post-processing and any corrections or QA/QC of the data.

Flight mission telemetry and trigger records | Spectral reflectance json files | Optical RGB photos | Spectral datacube image | Multispectral image | Multispectral calibration image | Calibration panel reflectance reference | Thermal IR images | Ground station |

Level 1 (L1): data products after basic post-processing

The Level 1 data represent the initial data products derived from the raw L0 measurements. These typically include the use of algorithms and approaches to combine individual measurements or images into a continuous image using approaches and GPS information to provide a georeferenced orthomosaic (i.e. GeoTIFF). Point cloud data from LiDAR systems will often be combined with UAS IMU/GPS positional information to create a single combined dataset referenced to the surface. For optical RGB orthomosaic data, the structure-from-motion derived digital surface model is also typically provided as a L1 product.

Optical RGB orthomosaic | Digital surface model | Thermal IR orthomosaic | At-surface spectral reflectance | At-surface spectral reflectance orthomosaic | Point clouds |

Level 2 (L2): higher-level processed data products

The Level 2 data represent higher-order processing of the L1 datasets to apply additional QA/QC, corrections or improved geo-registration prior to use in scientific analyses. In addition, L2 products will often include additional processing of height models to derive the vegetation height (canopy height model) and the estimation of the surface terrain without vegetation present (digital terrain model). Both L1 and L2 datasets should be provided with uncertainty estimates for geo-registration and any physical retrievals, at the pixel level or at least the image level (i.e. averaged over the entire orthomosaic). Filenames can include pixel size (grain size) information using the convention indicated, e.g. a 10 cm pixel can be represented as “point01m” in the file name.

Georeferenced optical RGB orthomosaic | Georeferenced thermal IR orthomosaic | Georeferenced optical spectral orthomosaic | Canopy height model | Digital elevation model |

Level 3 (L3): higher-level derived data products

The Level 3 data represent the highest-level of data processing and often utilize additional assumptions, approaches or other data to generate ecosystem data products. Examples include maps of land-cover / land-use, terrain feature mapping and characterization (e.g. polygonal tundra), or plant functional traits (e.g. foliar nitrogen). L3 datasets should include uncertainty estimates for any physical retrievals, at the pixel level or at least the image level (i.e. averaged over the entire orthomosaic).

Plant functional type (PFT) map | Foliar nitrogen |

Level 0 (L0): raw data collected with the platform

Flight mission telemetry and trigger records

--- Examples and comments
File type examples kml, txt, json
File name recommendation [date][flight number or record] ID][data_description]. e.g. 2018-07-26_flight1_camera_shutter_log.txt
Comments Depending on the UAS platform there may be ancillary information required for data processing. These records, including GPS information, trigger points, and other telemetry can be saved as L0 mission information.

Spectral reflectance json files

--- Examples and comments
File type examples json, txt, netcdf
File name recommendation [date][flight number or record ID][spectral_reflectance].txt
Comments If utilizing a point spectrometer system a flat text or other file will contain the spectral information by wavelength. L0 data should include the raw DN or radiance data and any in-flight or pre-flight calibration data needed for conversion to the measurement of interest (e.g. surface at-surface reflectance).

Optical RGB photos

--- Examples and comments
File type examples jpg, tiff
File name recommendation [prefix][number].jpg e.g.DSC0001.jpg
Comments Report the naming convention and total number of files/raw images. The individual file naming convention is typically set by the UAS camera.

Spectral datacube image

--- Examples and comments
File type examples dat, hdf5
File name recommendation [date][flight number or record ID][DN or rad or ref]_img.tiff
Comments Usually Digital Numbers (DN) swaths; file type is often in a proprietary or binary format. DN: digital number; rad: radiance image; ref: reflectance image.

Multispectral image

--- Examples and comments
File type examples tiff
File name recommendation [date][flight number or record ID][DN or rad or ref]_img.tiff
Comments Usually Digital Numbers (DN) swaths; file types are usually proprietary formats.

Multispectral calibration image

--- Examples and comments
File type examples tiff
File name recommendation [date][flight number or record ID][DN or rad or ref]_img.tiff
Comments Usually Digital Numbers (DN) swaths; file types are usually proprietary formats.

Calibration panel reflectance reference

--- Examples and comments
File type examples Txt, csv
File name recommendation [date][flight number or record ID][spectral_calibration].txt
Comments The pre-/post-flight measurement of the spectral calibration panel in DN or radiance.

Thermal IR images

--- Examples and comments
File type examples Binary, tiff, HDF5
File name recommendation [date][flight number or record ID][DN or temperature].dat
Comments Depending on the camera and image acquisition these data will be in a binary format or image format. Data should be provided with the information or assumptions necessary for converting raw radiance to brightness and surface temperatures.

Ground station

--- Examples and comments
File type examples csv, tex, rinex binary
File name recommendation [date][flight number or record ID][gps].txt
Comments GNSS ground station raw and post-processed data.

Level 1 (L1): data products after basic post-processing

Optical RGB orthomosaic

--- Examples and comments
File type examples tiff
File name recommendation ***_rgb.tiff
Comments An orthomosaic image that has been georeferenced to the surface with a coordinate system and projection.

Digital surface model

--- Examples and comments
File type examples tiff
File name recommendation ***_dsm.tiff
Comments An orthomosaic image that has been georeferenced to the surface with a coordinate system and projection. The DSM can either be derived from structure-from-motion or through the processing of LiDAR observations.

Thermal IR orthomosaic

--- Examples and comments
File type examples tiff, binary
File name recommendation ***_tir.tiff
Comments An orthomosaic image that has been georeferenced to the surface with a coordinate system and projection. Typically provided as a physical measurement such as land-surface temperature in degrees C or Kelvin.

At-surface spectral reflectance

--- Examples and comments
File type examples csv
File name recommendation ***_HDRF.csv
Comments Includes spectral/atmospheric calibration information.

At-surface spectral reflectance orthomosaic

--- Examples and comments
File type examples tiff, binary, hdf5
File name recommendation ***_HDRF.dat
Comments Calibrated and at-surface spectral reflectance image cube. If relevant, the data format should also be indicated in the metadata, e.g. band-sequential, BSQ or band interleaved by pixel, BIP.

Point clouds

--- Examples and comments
File type examples laz
File name recommendation ***_CloudPoints.las OR ***_CloudPoints.laz
Comments This should be the georeferenced, classified if possible, LiDAR or SfM cloud points.

Level 2 (L2): higher-level processed data products

Georeferenced optical RGB orthomosaic

--- Examples and comments
File type examples tiff
File name recommendation ***_GeoAligned_point01m.tiff
Comments Improved geo-registration of the L1 data product using an authority reference.

Georeferenced thermal IR orthomosaic

--- Examples and comments
File type examples tiff, binary, hdf5
File name recommendation ***_ GeoAligned _TIR_point01m.tiff
Comments Improved geo-registration of the L1 data product using an authority reference.

Georeferenced optical spectral orthomosaic

--- Examples and comments
File type examples tiff, binary, hdf5
File name recommendation ***_ GeoAligned _HDRF_point05m.tiff
Comments Improved geo-registration of the L1 data product using an authority reference. If relevant, the data format should also be indicated in the metadata, e.g. band-sequential, BSQ or band interleaved by pixel, BIP

Canopy height model

--- Examples and comments
File type examples tiff
File name recommendation ***_ GeoAligned _CHM_point01m.tiff
Comments The estimated canopy height is derived with the point cloud data from LiDAR or structure-from-motion and using an algorithm to detect the vegetation (if present).

Digital elevation model

--- Examples and comments
File type examples tiff
File name recommendation ***_ GeoAligned _DEM_point01m.tiff
Comments The estimated surface or terrain elevation derived with the point cloud data from LiDAR or structure-from-motion and removing the vegetation (if present).

Level 3 (L3): higher-level derived data products

Plant functional type (PFT) map

--- Examples and comments
File type examples tiff
File name recommendation ***_pft.tiff
Comments A map of plant functional types by pixel provided with an estimate of the error or uncertainty for each class.

Foliar nitrogen

--- Examples and comments
File type examples tiff
File name recommendation ***_leafN.tiff
Comments A map of estimated leaf nitrogen at the top of the canopy with an estimate of the error or uncertainty for each pixel or image.