The Herd

The Herd

AI-powered algorithms for Earth Observation

AI-powered algorithms to prepare and analyse Earth Observation data

The Herd is a set of AI-powered algorithms designed to facilitate Earth Observation data analysis. It comprises three elements – data preprocessing techniques, approaches to spatial resolution enhancement of image data, and data analysis algorithms.

The Herd fully supports the EO processing chain – from optical data acquisition, through data preparation and manipulation, up to data classification, segmentation, unmixing, compression, and much more.

Do you want to find out more

THE HERD IS BUILT OUT OF 3 KEY ELEMENTS

Data pre-processing

The algorithms designed for data preparation steps such as band misalignment correction, bad pixel masking, optical distortion correction, georeferencing, and sensor-level radiometric corrections.

Enhancement of spatial image resolution

The algorithms for single- and multiple-image super-resolution reconstruction utilise classical image processing, machine learning, and deep learning.

Data analysis

The algorithms, alongside rigorous validation pipelines, can unveil every single bit of information from your data through dimensionality reduction, feature extraction, classification, segmentation, hyperspectral unmixing, and time-series analysis using classical machine learning and deep learning.

All key elements of The Herd were designed to be used both on Earth and on-board the satellite – they fit FPGA processors and are compatible with the Xilinx-based Data Processing Units. The full power of The Herd can be unchained in high-dimensional data analysis – we handle hyperspectral, multispectral, and time-series data.

The Herd is part of the Smart Mission Ecosystem – hardware, software, and AI-powered algorithms designed to complete your mission.

KEY ADVANTAGES

Mission Cost Reduction


- Send only insights into the ground instead of huge volumes of raw data.

- Run ready-to-use algorithms to support your mission instead of developing them from scratch.

Mission Customisation


- Release the potential of data-driven algorithms.

- Make your mission multi-purpose or reconfigure it while in orbit.

Mission Simplicity


- Use The Herd as a mission building block and exploit only those algorithms which are really required for your task.

- Train the deep networks on the ground and uplink your fully functional and memory-efficient models straight up to the satellite.

Case study

The Herd will be used in the Intuition-1 mission, which is planned for launch in 2022.

Intuition-1 is a 6U-class satellite with a data processing unit – Leopard – enabling on-board data processing acquired via a hyperspectral instrument with spectral resolution in the range of visible and near-infrared light.

Contact us

Write, call us or meet us directly!