Define the information needed
What needs to be observed, understood or reported.
Making the physical world accessible to artificial intelligence.
01 — THE CHALLENGE
Artificial intelligence
needs access
to the real world.
Artificial intelligence can analyse, interpret and make decisions.
But to interact with the real world, it needs access to information about its physical surroundings.
A scene, a movement, a thermal signature: before an application can interpret them, they must be made accessible to it.
02 — OUR ANSWER: XPLOCAM
Your AI application.
Connected to the physical world.
A hardware platform that lets developers connect their AI applications to the physical world, without having to design the electronics and acquisition systems themselves.
Available now
For developers, integrators and researchers looking to experiment, build and prepare their perception applications.
XPLOCAM provides the hardware foundation. You develop the application that defines its purpose.
03 — OUR AMBITION: PERCEPTION INFRASTRUCTURE
Beyond a single device.
Perception infrastructure.
Designing the architectures and perception nodes that will enable artificial intelligence to be deployed in real-world environments.
On an industrial site, around a perimeter or within a mobile system, the architecture starts with local needs: what to perceive, process and transmit. XPLOCAM provides a foundation for exploring this approach.
Process close to the source
Produce information useful to the application, as close as possible to the field.
Control what is transmitted
Account for network conditions and adapt video quality to the available bandwidth.
04 — THE PLATFORM IN PRACTICE
Explore fused vision.
Build on a ready-to-use foundation.
For developers, integrators and researchers: access an integrated acquisition and processing chain, then focus on your application.
Acquire & fuse
Use visible imaging alone, thermal imaging alone or a fusion of both. The sensors are synchronised in hardware and processed through a shared pipeline.
Synchronised sensors · FPGA pipelineCode & experiment
Build perception, video analysis or embedded AI functions on NVIDIA Jetson, with a C++ API and ROS examples.
Linux · Jetson · C++ APIPrototype & integrate
Test your application in its intended environment, then prepare its integration with support and expertise from Tachysséma.
Documentation · Examples · SupportHardware paves the way for software.
You define the application. XPLOCAM provides the foundation to explore it.
VISIBLE AND THERMAL VISION — A REAL-WORLD TEST
Context from visible imaging.
Information from thermal imaging.
Live visible and thermal streams of the same scene, accessible from a remote workstation. On the left, thermal information is overlaid on the visible image in real time to highlight the subject. On the right, thermal imaging alone.
Real-time multisensor fusion and remote monitoring
Test carried out with XPLOCAM connected to a 5G router. Visible and thermal streams are broadcast live, while fusion parameters are controlled from a remote workstation approximately 20 km away.
One scene. Three views.
Visible only
Thermal only
Visible–thermal fusion
XPLOCAM · Test on 14 October 2026 · Successive captures, cropped without retouching
ARCHITECTURE & INTERFACES
From signal to system.
Embedded expertise.
Inside XPLOCAM: synchronised visible and thermal sensors, a CrossLink-NX FPGA pipeline and NVIDIA Jetson embedded computing.
The processing chain connects the sensors to the FPGA, then to Jetson and your applications. Video and network interfaces let you integrate this foundation into your environment.
01
Acquisition & signal
Combine visible and thermal sensors.
02
FPGA & real time
Synchronise the sensors and control the pipeline.
03
Jetson & software
Develop on Linux within the NVIDIA environment.
SENSORS & ACQUISITION
Visible and thermal
Sony IMX678 visible sensor, Lynred ATI640 thermal sensor, hardware synchronisation and MIPI CSI-2 acquisition.
INTEGRATION
Video and networking
Mini DisplayPort, USB-C and Gigabit Ethernet for displays, peripherals and communication.
DEVELOPMENT
Your software environment
Linux, a C++ API, ROS examples and the NVIDIA environment to develop your perception and AI functions.
APPLICATION AREAS TO EXPLORE
Different environments.
A shared need for useful perception.
Explore potential applications. Each project starts by defining what needs to be observed, under what conditions, and what needs to be transmitted.
POTENTIAL APPLICATION
Observe a site from the field.
For a perimeter or a remote infrastructure site, network availability and installation conditions are part of the perception requirements.
THE NEED TO DEFINE
What scene needs to be observed, and what video quality is required to use it remotely?
WHAT XPLOCAM BRINGS
Bring processing closer to the site being observed and adapt the transmitted video to the available connection.
These examples describe potential uses to explore with you. They are not deployment references.
OUR DIRECTION
Only the essential data.
All the richness of meaning.
Why transmit and store data before it has value?
Our approach: start with local needs and transmit only what is useful. XPLOCAM is part of this vision for perception infrastructure.
LET’S DISCUSS YOUR ENVIRONMENT
What would you like to make
accessible to your AI?
Looking to develop with XPLOCAM or explore perception infrastructure? Tell us about your application, your environment and the information you need.
YOUR CONTACT
Olivier Darracq
Tachysséma Développement
contact@tachyssema.frLet’s discuss your use case and the possibilities for integrating XPLOCAM.