My EOS Processing
Introduction to EOS Processing
EOS Processing – is a cloud-based tool for GIS specialists, analysts and business users to perform a remote sensing analytics. It’s implemented with a unique suite of analytic algorithms that extracts meaningful information, integrated into the family of EOS products that completes a full range of tasks in one suite and has a user-friendly interface.
1. The leading features of the instrument are:
- Imagery processing and analytics
- Neural Net-based algorithms
- Custom methods builder
- On-demand tasking
2. The instrument USP is based on:
- Online processing of huge remote sensing data
- Integration of neural networks
- Working with versatile sources of remote sensing data
- 3- Modeling solutions
- Integration of IPAs into third-party services
- Developing of own processing algorithms
Integration with EOS Platform
Being a part of an ecosystem of four mutually integrated EOS products, namely
- LandViewer allows even non-expert users to search, view, instantly process and download satellite multichannel images.
- EOS Storage is speciallly developed to store, share, distribute and get the instant access to thousands of Terabytes of Earth Observation and GIS related data.
- EOS Vision is employed to visualize data sets it creates data-driven maps with SQL-based interface and state-of-the-art cartographic tools.
Make a full circle of geodata workflow in one suit. Search, process on-the-fly and pick an appropriate image from the largest satellite imagery catalogue in LandViewer, save it to your personal (256Gb) cloud and get the instant access, share and distribute huge GIS related data in EOS Storage, extract meaningful information from the image with Net-based neural algorithms in EOS Processing and perform efficient analysis via visualization with SQL-based interface and state-of-the-art cartographic tools in EOS Vision.
In the context of EOS Processing, follow the scheme below:
The EOS Processing main functions are:
- Creating a project for big remote sensing data online processing
The instrument alllows to create a project for remote sensing data processing, raster and vector both. The data in the project is processed using the set of neural chains, available in the service.
To perform analytics the tool provides the specific operations for imaging data processing.
- Categorization of processing chains
It’s able to represent all the methods of remote sensing data processing in a certain structure, according to the category of operations performed and data sources (Landsat 8, Sentinel 2, SAR).
- Setting up processing algorithms
It allows to configure settings for Analytics&Calculations operations, when selecting a processing chain
- Working with files uploaded from EOS Storage
It’s able to load the results of EOS Storage file processing to the installed directory in this service.
- Data preview
It is able to search data in EOS Storage, using the set AOI and preview it on the map. The result file, both raster and vector, can also be viewed on the map.
- Creating custom processing chains
It’s possible to compile&run a unique algorithm for remote sensing data processing from accessible and implemented geospatial data processing operations
Terms and Definitions
Project – a folder that contains the dataset processing chain. The user is able to create several Projects at the same time to sort the data by the algorithms applied.
Workflow library – a section in EOS Processing to store all the algorithms and filters, available in EOS Processing. The user is able to create an individual algorithm here.
Workflow info – a section in EOS Processing, containing all the algorithm details, that is the description of the algorithm, the type of input&output data, the info about the satellites selected and bands combined to run the algorithm correctly.
Processing – a number of processes the user runs and gets the result according the algorithm, selected from the list, available in EOS Processing.
Settings – the settings to customise the algorithm before the processing starts. They are used to get more accurate results, introduce the user’s formulas and projections, choose approach, etc.
Add – a button, used to upload data from EOS Storage; once you click it the modal box with the list of data stored in EOS Storage shows up.
Band mapping – is a fixed list of bands and their binding to a satellite image. When the custom data is operated, the user set the Band mapping himself.
Palette – a set of options to work with the result data, that is the ability to change band combinations, apply palettes to single-channel images, the ability to work in grayscale mode
Band – a spectral channel in satellite image
High resolution band – a spectral channel in satellite image with a higher resolution capacity
Notification – a service, that allows to track user’s activity at EOS Platform, using end-to-end notifications.
AOI (Area of Interest) – the area, specified to run definite processes. e.g. search images by AOI
My AOI – a service, that allows the user to keep saved /downloaded custom AOIs
Process – the task, performed with applying of any of the algorithm
Labels – a button that serves to add attributive information to the map (roads, tags, etc.)
Search – a button that serves to search both within the Project and across all the created processes Projects, for algorithms to create new processes, etc
Prepare file for EOS Processing – During preprocessing, the image split into bands, each band on the picture is loaded separately into S3 Bucket for further work across EOS Platform products, meanwhile it builds a histogram, calculate the intensity range, etc.
Data formats – the types of data, currently used at EOS Processing:
- Raster file (GeoTiff or JP2)
- RAW data (Archive of geodata: tar.bz, tar.bz2)
- Vector file (Geojson, KML)
Fields in the settings:
- Calibration type – type of calibration (radiance, reflectance)
- Expression – a formula field in the settings of the Raster calculator to calculate indices
- Principal Analize – finding the main components in the input bands, specified by the user
- Shift – pixel shift image alignment
- Threshold – the process of сropping the input image below the threshold value with further binarization
- Morphology – a set of non-linear operations related to the shape or morphology of features in the image. Morphological operations apply the structuring element to input image and generate output image.
- Area filtering – the process of filtering the input image by area
- Resampling method – a method, applied to change sampling frequency (Cubic, Cubic, Spline, Lanczos, Average, Bilinear)
- Raster projection – a mathematical method to project a satellite image on a plane
- Iterations – the number of iterations to implement morphology
- Sigma (gauss filter) – the coefficient, used in “Gauss filter”
- Kernel size – the size of the kernel the operation is performed in
- Filter type – a selection of filters’ types for the processing (Sobel Median Laplace Gauss)