ESA title

Artificial Data Enhancement - in1

Data Segment
  • Data Processing & Visualisation
Does your satellite data suffer from sub-optimal resolution and haze? High resolution data is expensive, may lack temporal and geospatial coverage and is not accessible to everyone. These problems can be overcome with in1, which performs Artificial Data Enhancement (ADE). The in1 toolkit can be used to enhance your satellite data in real-time and at low cost, improving its usability and unlocking exciting new use cases.
Objectives of the Product

High resolution satellite data has enormous potential, but is expensive and not widely available. With limited budget and resources one is often left with data that suffers from haze, distorted colours, and sub-optimal resolution. These problems severely limit the usability of the data.

in1, utilising Artificial Data Enhancement (ADE), provides a solution. Based on state-of-the art software technology, the in1 toolkit can be used to enhance low value satellite data at low costs. Due to the increased resolution, the enhanced data has improved usability and unlocks exciting new use cases for existing datasets and satellite programmes. For example, users will be able to perform more accurate data analyses.

The in1 tools are offered as part of an application programming interface (API) that allows for real-time data enhancement. The API is compatible with different programming languages and will become available as on-premises software in the future, enabling unlimited use.

The image below illustrates the enhancement process from low- to high-resolution data using a few lines of coding for the API.

Customers and their Needs

in1 responds to the needs of three types of customers:

  • Data providers: with in1, satellite data providers and resellers are able to perform data enhancement on large volumes of data. This increases the impact on their user communities. Data providers can offer their customers enhanced data as a standard product or provide the choice to apply ADE algorithms to the provider’s raw data. For data providers, it is important for in1 to be able to process large volumes of data in a short amount of time.
  • Analytics providers: better data leads to better analytics. Analytics using big data is growing rapidly due to the availability of on-demand computing power, data storage and more satellite data. With in1 they have low-cost and real-time access to high-value data products in large quantities, improving the quality of the analytics. It is important for in1 to be easily integratable into the applications and data pipelines.
  • Value adding service providers: companies and organisations which focus on platform-enabled value-adding services for security, agriculture, energy or otherwise, can increase the value of the platform. Using in1 their customers can obtain more useful insights. The operations of these platforms and tools require high throughput without latency in any region of interest.

The image below illustrates the product positioning:

Targeted customer/users countries

The in1 solution targets customers worldwide with a focus on the European Union, Canada and the United States.

Product description

in1 provides a solution to problems with satellite data, such as haze, distorted colours, and low resolution. Based on state-of-the art software technology, the in1 toolkit can be used to improve the value of satellite data. The artificially enhanced data is carefully validated, ensuring that it has the characteristics of measurement data and that the physics governing the acquisition process is preserved.

The in1 tools are offered as part of an API that allows for real-time data enhancement. The in1 architecture makes use of the most recent scalable infrastructures that enable on-demand use. The API is compatible with different programming languages and the processing module will also be available as on-premises software in the future, enabling unlimited use.

Added Value

The benefits of enhanced data quality are:

  • Improved spatial resolution, which allows for new types of use cases that were previously impeded by limited resolution.
  • An increased number of usable products or data points due to the removal of haze. This means more frequent observations with an increase of up to 20%.
  • A competitive edge in the market due to the delivery of higher-value data as part of a platform or service.
  • Saved costs by using relatively cheap data that is brought to the level of higher-priced data (e.g. replace Spot 6-7 data by Sentinel-2 data).
  • Decreased processing costs due to not having to run computationally intensive and maintenance-heavy haze removal algorithms.

Current Status

The ADE activities have been concluded mid-July 2021. Our marketing and outreach efforts gave us unique insights into the EO market dynamics. As a result we have decided to look for partnerships with EO data providers such as satellite operators rather than offering in1 as a service to anyone working with EO data, because the time to market is too long for us. We developed relevant metrics and quality assurance parameters for the AI generated images to give users the confidence that the highly appreciated radiometric accuracy of the Sentinel-2 BRGI are preserved, whilst enhancing the pixel resolution of the original Sentinel-2 images by 4.

B and D represent enhancement and fidelity respectively of the in1 super-resolved Sentinel-2 visible bands in sub-figure A (left is bilinear upsampling, right is in1 super-resolved). The results are further corroborated by the spectral value histograms (C), whereby the distributions follow similar trends. The spectral histogram displays the frequency distribution of pixel values per spectral band for both the original Sentinel-2 product and the one processed with in1. Values have been normalized between 0 and 1 for comparison purposes. Slight variations are observed for urban landscapes due to the enhancement of the lines around objects. 
Prime Contractor Company
Sobolt B.V.
Netherlands Flag Netherlands
Contractor Project Manager
Danny Hetharia
Weena 592-594, 3012 CN Rotterdam

ESA Technical Officer
Patrick Griffiths
Related links
in1 website arw

Current activities