Stream, Analyze, and Query Hyperspectral Imagery in Real-Time
  • Home
  • Technology
  • Industries
  • About
  • Contact
  • Blog
The biggest challenge of multispectral and hyperspectral imaging is handling and processing the sheer amount of raw data. Hyperspectral images are volumetric image cubes that consist of hundreds of spatial images at particular wavelengths.

Technology

Picture
Superior compression capability

Metaspectral builds on years of in-house research in spectral data compression capability, as well as numerous work by organizations such as the MPEG consortium and the Consultative Committee for Space Data Systems (CCSDS).

CCSDS is a consortium of national space agencies, including NASA and ESA. Their newly formalized standard CCSDS-123 directly addresses the lossless and near-lossless compression of multispectral and hyperspectral data.
​
Our proprietary implementation, inspired by space applications, supercharges this standard to bring unparalleled compression throughput of up to 37 Gbps while only consuming 10% of the compute fabric compared to other alternatives.
​
Performance​

By losslessly compressing spectral data, achieve up to 67% reduction (33% of the original size). Near-lossless compression, on the other hand, achieves up to 90% size reduction of your data.

AI Analytics​

Machine Learning and AI is in our DNA. Our algorithms process pixel-level spectral IDs at unparalleled accuracy for classification, and are easily trainable to recognize any type of underlying material, whether detecting Improvised Explosive Devices (IEDs) in defense applications, or sorting between different types of plastic bottles for the recycling industry.

Indexing and Storage

Transform unstructured data into structured, searchable dataset. Data is indexed according to classes extracted during analysis, and made available for querying in real-time.

MLVX Technologies Inc. 2020.
  • Home
  • Technology
  • Industries
  • About
  • Contact
  • Blog