Why is Resolution so Important?
Automotive Engineering
October 8, 2022Calibration Dilemmas
October 7, 2022
When it comes to infrared, “if you can’t see it, you can’t measure it”
This is a key point because whether we’re engineers checking bearings or electrical inspectors checking for bad connections, an inspection that is blind to a fault is worse than no inspection, as it can lead to a false sense of security.
Glossary
Field Of View and Instantaneous Field of View
The Field of View (FOV) is simply a measure of how much your camera can see. It is similar to the width of a car windscreen. If it’s a wide windscreen, it has a wide FOV.
Instantaneous FOV (iFOV) asks us, what is the smallest object we can see through our windscreen?
The epitome of domestic car engines is the V8. It’s balanced, it’s powerful and it’s ‘root note’ (see the article on resonance) is throaty and deep.
Ford and General Motors approach the V8 quite differently. The LS engine is the standard V8 for most General Motors Vehicles (Chevrolet Corvette (see example below), Holden SS, HSV etc)
The Ford Coyote is a much more modern design and yet the LS engine of the Holden/GM regains fabulous and popular.
In this picture, we see a typical car windscreen with a wide FOV (in this case it is about 125 degrees). Through his screen, the driver can just make out the sign warning of wandering cattle, some meters ahead. Anything smaller than this he can’t see.
The sign is the driver’s iFOV. The smallest object he can define.
When measuring temperatures with Infrared (whether a thermometer or a camera), what we are measuring is NOT defined by the helpful laser spot marker equipped on most devices.
The laser merely marks the very (sometimes very, very) inaccurate center of a measurement spot, whose size varies when we get farther from our target.
With an IR thermometer, it’s fairly simple. We have a simple distance-to-spot ratio as shown in the diagram below. The further away we get from the target the bigger the measurement area becomes.
If the measurement spot is bigger than the target, the temperature displayed will have massive errors! In fact as the measurement spot gets larger than the target, the errors grow exponentially!
So it is mission critical that the measurement spot MUST be smaller than the target. An IR thermometer is a thermal camera with a 1-pixel by 1-pixel resolution. A thermal camera has many pixels, sometimes several thousand!
Because you can’t be sure that the fault image will fall squarely on a single pixel receptor, one needs more than 1 pixel within the measurement spot. There is a generally accepted idea in the IR Inspection world that:
To see a fault, one needs a measurement spot containing 3-pixels
To measure a fault, one needs a measurement spot containing 5-pixels
So the mathematics gets a little more complex, but don’t worry! Homershams have constructed a spreadsheet to do these calculations for you.
Science not Sales!
Thermal Imaging cameras, range from $2,500 to well over $80,000, so clearly there are some big differences between models. How can the potential user select the appropriate model?
The single most important factor in selecting a thermal imaging camera is TARGET SIZE. “How big is the object I’m wanting to measure and how far away from it am I?”
In one fell swoop, the answer to this question reduces ones short list from tens of models to a handful.
The danger of not understanding target size vs. model selection is that one may choose a camera that either misreports what the temperature is or doesn’t even detect a fault! In this case it is worse than having no Infrared inspection at all, as one might be lulled into believing all is well, where in actuality, faults exist.
Now this danger exists both for close-up work (e.g. switchboard inspection) as well as distant work (e.g. Power Line Maintenance).
For example a reader of this article might be surprised to read that perhaps NZ most popular camera, is NOT good enough for many switchboard inspection jobs.
The good news is that checking the target size requirements is “science not sales!” One can reasonably simply calculate the required specification for any given target size/distance using trigonometry, but as mentioned above Homershams have helpfully done this on a spreadsheet to make it easier.
- A copy of the spreadsheet can be downloaded from Homershams website by clicking here
(It is generally accepted by the industry that to detect a fault one needs a spot size, containing 3-pixels, that is smaller than the target and that 5-pixels are required to measure the actual temperature.)
Here is a theoretical example: As mentioned a very popular, high-end brand, IR camera sold in NZ in good numbers, has a minimum 5-pixel spot size of 11 mm at its closest focus distance. 11 mm is just too big if you consider electrical cables of 3 mm diameter. One might easily completely miss a fault and only become aware of it, when the Fire Brigade or the customers insurance company contact you!
Here is a real example:
One of Homershams sales team was delighted to discover a line fault in Kaianga Road, Christchurch during 2012.
When walking out of a customer’s premises, they looked up at a power pole with their IR camera.
Here is the image that was captured.
It was a cold day (around 10 °C) and one can see in the image that there is a serious fault on 1 of the phases. 86 °C is very hot for those conductors!
Now this was a good camera, but not quite good enough for this situation. He was perhaps 11 metres from the joint and the camera had 320 X 240 pixel resolution and 21 degree FOV. So it had a 5 pixel measurement spot of around 60 mm.
The staff member went back with a higher resolution camera, with a lens of tighter FOV and got this image.
This camera has a 5-pixel spot size of around 20 mm at the same distance. Note how the temperature is now reported as 58 °C (not 85 °C). This WAS the actual temperature.
This was a classic example of spot size to target size error. Sure both cameras showed the fault, but the first one had an error of over 35 % and although in this case, over-reported a fault, could have just as easily under-reported it.
The fault turned out to be a bad connection on an LV-Clamp. (The clamp that connects to the 11kv/22kv main line and feeds down to the transformer)
This image is the offending LV-Clamp with the measurement spot sizes from the 2 cameras superimposed to demonstrate.
***STOP PRESS!*** August 2014
Homershams are delighted to announce their appointment as the NZ distributor for ULIRvision IR cameras. This adds to their range of cameras which includes Optris and Jenoptik.
It’s something of a secret amongst IR camera suppliers, but there are only a handful of manufacturers of the IR sensor and of the IR lenses, and some are better than others.
The sensor used in the ULIRvision camera is from a French company ULIS and it’s renowned as a particularly good one.
Typical camera resolutions are as follows
120 X 120 pixels
160 X 120 pixels
320 X 240 pixels
384 X 288 pixels
640 X 480 pixels
1024 X 768 pixels
2048 X 1536 pixels
320 X 240 is a very common size for Preventative Maintenance and Electrical Inspection for the reasons mentioned in the article above.
But moving to 384 X 288 pixels gives a significant increase in quality, greater than the simple 20% increase in pixel count, due to the excellent ULIS sensor.
(Some of Homershams Optris cameras use the same sensor)
It’s something of a secret amongst IR camera suppliers, but there are only a handful of manufacturers of the IR sensor and of the IR lenses, and some are better than others.
The sensor used in the ULIRvision camera is from a French company ULIS and it’s renowned as a particularly good one.
ULIRvision manufacture cameras from 160 X 120 through 640 X 240 pixels, but for the reasons mentioned in the article above, Homershams will only supply cameras of resolutions greater than 320 X 240 unless a customer has a very specific reason to use a lower resolution.
Homershams have determined that the best model to stock is the TI395.
Homershams invite you to take the challenge and compare the ULIRvision camera with the other excellent cameras in its class.