Real-Time Imagery Service

Access your data anywhere, anytime, and at any scale.

Coming Spring/Summer 2025

Manage Petabyte Imagery Datasets in Real-Time.

Thanks to advances in Wide-Area Motion-Imagery (WAMI), it is common to build high-resolution time-series imagery of entire cities in real-time. With our NITF and MIE4NITF technology, we are building the future of imagery management. Our Imagery Service will allow users to actively write imagery to your local filesystem while imagery is being served to viewers.

We aim for an alpha-release of our Imagery-Service in Spring or Summer of 2025. If you are interested, please email us for more information.

What is RTIS?

Workflow From Sensor Integrator’s View into RTIS

Companies who build and/or integrate sensors have a delicate balancing act. There is a need to produce a large volume of data, with precise telemetry and projections, with adequate compression, and a robust means of distributing and offload data. Terminus can assist with this process. Our Real-Time Imagery Service expects constructed NITF or GeoTIFFs. Our NITF and Geospatial SDKs can help you achieve the rest. We provide libraries for generating RPCs for your own camera model and abstractions to help encode and process J2K codestreams.

RTIS Core Capabilities

A core value of RTIS is that we do not modify your imagery when saved to disk. We will reproject only when requested. This provides maximum flexibility and enables you to retrieve whatever data you want in real-time.

Planned outputs include the original data, reprojected frames, and perspective or ortho-rectified motion imagery and video.

Why Terminus

Open Standards

Your data remains yours. Our service will store your imagery in the same manner it arrived. All data products use either existing Open-Source or publicly available formats. If we decide to use intermediate formats, we will publish an open-source specification for it. We will also publish open-source command-line tools for managing and converting data.

No Need to Reproject

With our Imagery-Service, you can provide raw or unprojected imagery with the required projection information. We will project on-the-fly. For NITF and GeoTiff users, this can be via a projection (transform or MAPLOB/GEOLOB), Ground Control Points (GCPs) or through Rational Polynomial Coefficients (RPCs). For best performance, we recommend providing projected imagery.

Operational-Ready

Our interface is designed for use in a data center, on-board an aerial platform, and anything in-between. We are designing our Imagery-Service to support write operations while serving imagery to downstream users. Our projection logic uses Vulkan Compute Shaders allowing for GPGPU performance without NVIDIA lock-in.

Supported Data Products

  • Still Images

    Request raw/original image frames (as sent by the sensor) as well as reprojected imagery (Ortho/Perspective) suited to your coordinate system. Supported formats are planned to be NITF, GeoTiff, JPEG/PNG/other OGC formats.

  • Motion Imagery

    NITF frames with MIE4NITF TREs served in a Zip, single image, or binary format depending on data.

  • Full Motion Video

    Planned support for streaming MPEG formatted video. Due to complexity of the NGA's KLV format, we will provide video first with KLV at a later date.

  • Metadata / Scene Info

    Access projection information, geographic bounds, time series data, security information, and anything stored with the original imagery.

  • Command and Control

    Our interfaces will support adding and removing imagery, collection operations, and other essential tasks via our published interfaces.

Planned Interfaces

REST API

Standard Imagery and Map Information

OGC-compliant Web-Map-Service (WMS)

OGC-compliant Wide-Area-Motion-Imagery (WAMI)

Terminus IPC API

Current Interfaces

  • ZeroMQ

Planned Interfaces

  • mvapic MPI

Benefits

  • All types work with Protobuf for easy (de-) serialization.

  • Persistent connections

  • Higher Bandwidth

Shared Library Interface

Benefits

  • Your application can send and receive data from the Imagery-Service directly from our shared library.