Some comments regarding the video…..
1. FLIR do make high speed thermal cameras. My SC4000 cooled MWIR camera is capable of 43K (43000) FPS !
2. FLIR bought CEDIP to increase their range of science and high end cameras
https://www.vision-systems.com/cameras-accessories/article/16747544/flir-completes-acquisition-of-cedip-infrared-systems3. The SC4000 camera and its progeny all come from the original Amber Radiance and Indigo Phoenix design and development streams. Some of FLIRs new offerings are CEDIP cameras with FLIR branding. The Infratec is just a rebadged CEDIP design as Infratec like to take other manufacturers products and stick their branding on it. This can be significant if seeking support. I needed some assistance from FLIR for my SC4000 camera and they provided the original documentation plus software because they made it. It is not the same when asking a company like Infratec for help. Infratec rebranded some Jenoptik IR-TCM384 thermal cameras that I have and they could not have been more unhelpful (unfriendly) when I asked about user manuals and software. Jenoptik and their sales agent were far more friendly and helped me with the user manual.
4. The commentator stated that he could not find much information on the camera or it’s user manual. The same is true of the FLIR SC4000. The reason is the advanced capabilities of these science grade high speed cameras. The SC4000 was designed for military applications on missile firing ranges ! It was released to approved commercial and scientific users but most technical information about the cameras was, and still is, heavily controlled. FLIR will only provide the user manual to a pre approved customer as it contains deep design and operation detail far beyond what would normally be found in a generic commercial thermal camera manual. The cameras also offer the user far more control over the sensor array and system settings than any general purpose camera so they are very versatile. Later, less capable versions of my camera do not have the same level of sensitivity and their user manuals may be downloaded from FLIR. Even the earlier Indigo Phoenix user manuals remain controlled release though.
5. When asking for help with my SC4000 configuration, I was put in touch with FLIRs scientists in Paris, France as they deal with the cooled camera support. I believe CEDIP had a base in Paris, France as well…. Coincidence or has FLIR moved cooled science camera support to the CEDIP offices ? I might add that talking to a very friendly senior scientist at that establishment was amazing. He could not have been more helpful and liaised between FLIR USA and me when it came to providing documentation and software that was long obsolete. He found what I needed and provided it and much other useful information without hesitation. He and the USA based support team had no access to an SC4000 or SC6000 as they are obsolete. The USA technical support team panicked when I sent them the cameras full internal configuration flash contents as they thought I had hacked the camera ! I explained that the SC4000 and SC6000 download a file containing all that configuration data to the host when the camera is connected to a fresh installation of the Researcher Software. They were shocked because later releases of the camera prevent user access to that configuration data as a deliberate act by FLIR to prevent user changes to the cameras capabilities ! I did say the SC4000 was heavily controlled and this is one reason why.
6. Dead pixels are a fact of life with thermal sensor FPAs of any significant pixel count. Just as with microbolometer arrays, the cooled cameras identify the dead pixels and hide them from the user using output values from surrounding pixels. Top of the line cooled science cameras have fewer dead pixels than budget uncooled cameras but they still have a few to be identified and hidden. The camera is tested at the factory to identify all dead pixels and correct them inside the cameras processing stages. It is also possible for the user of a science grade camera to make their own dead pixel maps on the host PC and use that instead of the built in dead pixel map. On some science grade cooled cameras the user can create a new dead pixel map on the camera itself during the user NUC setup process. This updates the map held in the cameras flash memory so is persistent after a reboot.
7. Calibration of science grade cooled thermal cameras…. These high end cameras are very different to general purpose uncooled thermal cameras as was shown in the video. They are far more capable but also require more configuration, at least upon initial use. In the case of my SC4000, the camera came with two calibrated ranges that were setup at the factory. There is room on the cameras calibration flash memory for more but FLIR charge ‘per calibration run’ so you tend to purchase only what you need for these cameras at time of purchase. The calibration run is carried out with many settings in the camera fine tuned for a specific task or role. This makes these camera excellent performers within the calibration range parameters, such as temperature coverage, frame rate, integration time, FPA biases etc. The difference with the SC4000 and some other science grade cameras is that FLIR is willing to provide user calibration options in the form off a special configuration GUI. This GUI was what I asked FLIR for and they provided after confirming I was a suitable recipient. That GUI is controlled release by FLIR and was only made available to some customers, and only upon request. That GUI is capable of making changes that can really mess up the cameras imagery, so the FLIR position on the release of that GUI is somewhat understandable.
I worked out the password for the hidden ‘factory use only’ menus in that GUI and now have access to all the configuration settings that FLIR use at the factory. This capability is almost certainly not provided on the later versions of the camera as FLIR want the camera sent only to them to add calibration tables etc. It is an income stream after all.
8. The Calibration process ? The cooled science cameras measure thermal energy differently to a microbolometer based camera and some terminology is used that will likely be unfamiliar to many users of uncooled thermal cameras. The pixels are Photon detectors and these provide an output to the high speed ADC and produce ADC COUNTS. These counts need to be ‘translated’ into Temperature units using calibration curves for use by many users. With these science cameras, an uncalibrated configuration will produce an image but the ‘measurement’ of pixels provides COUNTS and not temperature units. The camera is calibrated using either it’s internal calibration flags (if fitted) or external Black Body thermal references. The external references can be a more accurate approach. The SC4000 contains two temperature reference flags whilst other cameras may just use one, or have none. A calibration routine is run with all the FPA configuration data entered and a specific lens fitted. An external reference calibration includes the lens in the calibration but the internal reference calibration does not. Lens calibration is available though. The calibration routine checks for Dead Pixels, creates an NUC table and saves all required data into flash memory or to the host PC. There is no ‘flat field correction’ as such on a cooled camera as the NUC table captures inter-pixel variances and corrects them whilst operating at 77K. There is no significant pixel temperature drift as the FPA is temperature stabilised so no need for further corrections unless the user decides to instigate a NUC event for a fresh NUC table.
Some cooled thermal cameras, like the Amber Radiance, required the creation of a fresh NUC table upon each use in order to ensure an accurate flat field. In that case an internal NUC temperature reference flag produced the scene for a fresh NUC table in a matter of minutes. These cameras effectively need a babysitter when being set up. They are not normally “Switch on and go” straight out of the box, but some can do that. Once configured and calibrated to the users needs, the camera is pretty simple to drive and use.
9. Image analysis software ….. high end science cameras tend to use high end (read very expensive) image analysis and camera control software. This is just part of the purchase price to a buyer who can afford such technology. Sadly the software is often missing from used cooled science camera purchases. Many companies have software Governance in place so it is lodged elsewhere from the camera and only installed on the host PC in accordance with the EULA and any hardware licence dongle used. Without the correct software a cooled science camera can range from an ‘Expensive Doorstop’ to a fully self contained and independent system capable of use without a PC. Fully independent types of camera have their own keypad for configuration of settings but advanced settings may still require a host PC and specialist software. Many high speed thermal cameras provide a dedicated data output for the high speed data, be it SCSI, Cameralink, HOTLink, RS422 multi channel, or Gigabit Ethernet. My SC4000 uses Gigabit Ethernet and CameraLink. The USB 2.0 port is mainly for configuration and is not a high speed video output. Many high speed thermal camera systems also have an optional high speed data host to capture data on a dedicated image repository for later analysis. This is often the highest performance solution in terms of frames per second as it is tuned to the needs of the camera. Some cameras are basically a dumb data collector that has to be fully remote controlled with no autonomy and no ability to be used as a thermal camera without a suitable host computer equipped with a bespoke data link card and bespoke software. Those cameras are best avoided unless the host PC and software is included. The camera detailed in the video appears similar to my SC4000 in that it uses a PC Host for its configuration but it does offer a thermal image output even without a computer connected using a ‘last known’ configuration. Sadly without the correct software on a host computer, the camera may present a challenge when it comes to calibration and configuration. The command sets for these cameras is not always in the public domain, as previously stated. Manufacturers of this kit are not always keen to assist 2nd hand users due to the equipments sensitivity and age. As such it is not uncommon for a manufacturer to decline supply of user manuals, software, utilities or even the connector pin-outs. These are very much a specialist product and support can be extremely hard to obtain. Read my journey with the Amber Raytheon Radiance 1 cooled camera to see how ai managed to obtain user manuals and software fir that elderly camera….. spoiler alert… I tracked down the cameras senior design engineer via Facebook !
https://www.eevblog.com/forum/thermal-imaging/the-story-of-a-radiance-1-camera-and-frasers-quest-to-find-information-on-it/Well that is enough from me.
Fraser