This high energy Physics research organisation needed to acquire and store thermal images of their experiments. The experiments have a very short duration, but the time that they begin is not very controllable or predictable.
They had already identified and selected an Infra Red (IR) camera that operates at up to 1000 frames per second. This allows them to really capture the detail in the experiment, but generates a huge quantity of data to store. The camera also offers a gain control which allows it to image the whole range of temperatures that occur during the experiment, but this had to be controlled by some external signals.
The equipment used in the experiment uses high voltages and magnetic fields, so the system needs to be carefully designed to cope with that harsh environment and be safe to use.
The camera is mounted in the high magnetic field and has an analogue video output. We were concerned about the noise that might be introduced to the cabling of the analogue video by the changes in that magnetic field, so we wanted to digitise it close to the camera, and to transmit the data over optical fibers. We produced a test board for that, and trialled it in the magnetic fields of the real machine to be certain that there were no issues.
Having successfully developed a system to digitise and transmit the video, the rest of the system was placed outside of the magnetic field and connected only using optical fibers to ensure electrical safety. There the video was fed to a Digital Signal Processor (DSP) that could be programmed to find the hot and cold spots in the image on a frame by frame basis. This information was used to determine the gain setting that was applied to the camera for the next frame.
The DSP was also able to implement user defined thresholds to detect the start and end of the experiment results. This was used to control the storage of the data, along with timestamps of when it started and finished. A user defined mask was also used to select which sub-set of the image would be stored. Using these techniques the system minimises the amount of data that needs to be stored, while retaining all of the information needed.
The data to be stored was then buffered in the DSP memory, and written out to a network disk via Ethernet.
This system was put together using some off the shelf boards, connected to some custom PCBs. We combined these with an Industrial PC, and programmed the system for them.
After installing the system and using it for some time, they ordered a second identical system to use.
Then, having used the systems for more than 10 years, one of the cameras failed. They were unable to obtain an identical replacement camera, so asked us to look at making the systems interface with a new camera type. When we looked at the work required to change the systems, it became clear that it would be cheaper to supply some replacement systems using modern technology, so we proposed that to them. In the ten years since producing the original systems, electronics had moved a long way, and with the expertise we had from making the original systems, we could quickly program the new systems to provide a more modern and more capable replacement.