RemoteSensingToolbox

RemoteSensingToolbox is a pure Julia package built on top of Rasters.jl for reading, visualizing, and processing remotely sensed imagery. Users may refer to the Tutorials section for examples on how to use this package.

Installation

To install this package, first start the Julia REPL and then open the package manager by typing ]. You can then download RemoteSensingToolbox directly from the official Julia repository like so:

(@v1.9) pkg> add RemoteSensingToolbox

Once RemoteSensingToolbox has been installed, you can import it like any other Julia package. Please note that many features require you to also import the Rasters package.

using RemoteSensingToolbox, Rasters

Features

This package is a work in progress, which means that new features are being added and existing features are subject to change. To contribute, please create an issue on GitHub or open a pull request. A summary of both existing and planned features is provided below:

Feature	Description	Implemented
Reading and Writing	Read layers from a scene and the write results to disk	Yes
Visualization	Visualize images with various band composites	Yes
Land Cover Indices	Calculate indices such as MNDWI and NDVI	Yes
QA and SCL Decoding	Decode Quality Assurance and Scene Classification masks	Yes
Pixel Masking	Mask pixels to remove objects such as clouds or shadows	Yes
PCA	Perform PCA analysis, transformation, and reconstruction	Yes
MNF	Minimum Noise Fraction transformation and reconstruction	Yes
Signature Analysis	Visualize spectral signatures for different land cover types	Yes
Land Cover Classification	Exposes an `MLJ` interface for classifying land cover types	No
Endmember Extraction	Extract spectral endmembers from an image	No
Spectral Unmixing	Perform spectral unmixing under a given endmember library	No

Rasters.jl

RemoteSensingToolbox is intended to be used in conjunction with the wider Julia ecosystem and as such, seeks to avoid duplicating functinalities provided by other packages. As the majority of methods accept and return AbstractRaster or AbstractRasterStack objects, users should be able to call methods from Rasters.jl at any point in the processing pipeline. A summary of common functionalities offered by Rasters is provided below:

Method	Description
`mosaic`	Join rasters covering different extents into a single array or file.
`crop`	Shrink objects to specific dimension sizes or the extent of another object.
`extend`	Extend objects to specific dimension sizes or the extent of another object.
`trim`	Trims areas of missing values for arrays and across stack layers.
`resample`	Resample data to a different size and projection, or snap to another object.
`mask`	Mask a raster by a polygon or the non-missing values of another Raster.
`replace_missing`	Replace all missing values in a raster and update missingval.
`extract`	Extract raster values from points or geometries.
`zonal`	Calculate zonal statistics for a raster masked by geometries.

Satellites

When working with remotely sensed images, the user is often required to adapt their approach according to the type of sensor that produced it. For example, decoding digital numbers into radiance, reflectance, or temperature requires knowledge about the encoding scheme used by the satellite in question. To address these issues, we provide the AbstractSatellite type, which encodes various sensor-specific parameters at the type level for several popular platforms. This approach allows the end-user to write generic code that automatically adapts to the sensor being used. For example, if we have bound the variable src to an instance of Landsat8, then we can load a cloud mask from the included QA file by calling Raster(src, :clouds), or calculate the MNDWI index by calling mndwi(src).

SatelliteDataSources.AbstractSatellite — Type

The super-type of all satellites.

Sub-types must implement the AbstractSatellite interface.