The Vision4ce CHARM150AGX from Chess Dynamics.

Chess Dynamics unveils CHARM150AGX, a new AI video processor

The Vision4ce CHARM150AGX, the latest video processing module from Chess Dynamics was announced in a July 31 press release. The new hardware is designed for direct integration with video sensors and incorporates the NVIDIA Jetson AGX Orin AI engine. This development aims to enhance video analysis and decision-making capabilities at the operational edge.

The CHARM150AGX expands upon the existing CHARM product line, offering twice the processing capacity of its predecessor, the CHARM100NX. “It is optimised for video processing with support for standard video interfaces for direct connection to cameras. It also includes an FPGA processor and video outputs to support low latency video processing,” Steve Hogg, Director of Image Processing at Chess Dynamics told Calibre Defence via email. 

The increased processing power of the new processor allows for the simultaneous processing of data from multiple sensors, such as thermal and visible-light cameras, which facilitates a move toward multi-sensor fusion, the press release states. The system is ruggedised to operate in challenging environmental conditions, supporting applications in defense, surveillance, and autonomous systems.

The NVIDIA Jetson AGX Orin AI engine is a powerful, compact computer specifically designed to run artificial intelligence tasks in the real world, outside of a data center. Think of it as a small, rugged supercomputer for AI.

Its main purpose is to process data from various sensors—like cameras, microphones, and other detectors—right where the action is happening. This is known as “edge AI.” Instead of sending all the raw data back to a central server or the cloud for analysis, the Jetson AGX Orin can make smart decisions instantly and on the spot. This is critical for applications like autonomous robots, smart surveillance cameras, and self-driving vehicles where every millisecond counts.

Essentially, it’s a specialised chip that combines a standard computer processor with a powerful AI processor, allowing it to perform complex AI tasks while being small enough to fit into a wide range of devices. In the CHARM150AGX “it is designed to host the Vision4ce FrameWorkx image processing software, which builds upon the field proven detection and acquisition with robust tracking (DART) video tracking software,” the press release explains. 

Both Vision4ce FrameWorkx and the DART video tracking software are products from Chess Dynamics. Vision4ce FrameWorkx includes various tools for tasks like object detection, tracking, classification, image stabilisation, and multi-sensor fusion. DART is one algorithm that handles object tracking within the software suite. Other AI-powered algorithms in the Chess Dynamics range include the proprietary DEFT (Deep Embedded Feature Tracking) algorithm, the company’s website explains. DEFT is an algorithm improved through reinforcement learning to offer more robust target tracking in cluttered environments, and for targets that are mobile. 

Calibre comment

It is possible to use the NVIDIA Jetson family as well as Raspberry Pi chipset as a standalone video processor for some AI applications, providing a relatively cost-effective edge computing solution. Jetsons are understood to be used by companies like Helsing, and FlareBright has employed the Raspberry Pi for some of its latest applications. However, defence often requires more ruggedised and resilient systems than would normally be given to a commercial component. Ruggedising a component will typically increase its size, weight, cost, and potentially its power consumption if cooling is required, which will generally require specialised knowledge and engineering as well as awareness of defence requirements. It appears that the Chess Dynamics offering is aimed at this market, providing new video processing solutions that meet the emerging need for target recognition and detection on the edge to process data from a variety of sensors. 

By Sam Cranny-Evans, published on August 6, 2025. Credit for the lead image is Chess Dynamics.  

 

Get insider news, tips, and updates. No spam, just the good stuff!