Results 1 to 1 of 1

Thread: Image Fusion, Blending, Rendering, Multi Spectral Imaging...

  1. #1

    Image Fusion, Blending, Rendering, Multi Spectral Imaging...

    I've lost a lot of my life to answering questions about Image Fusion, Image blending, Visual Rendering and Multi Spectral Imaging.

    I am talking about all the graphical gimmickry that manufacturers perform by blending visual and thermal images into a single image.

    Whatever they call it, it doesn't impress me. In my opinion it’s a great party trick but has little useful application in real world thermography (except it's good for selling cameras).

    Image fusion (or any similar blending function) has been a very successful marketing gimmick and it’s no wonder the manufacturers persist with this function because it sells cameras like hot cakes.

    My main issue... distracting as much as adding
    When visual aspects of the image are integrated into the thermal image you are interfering with the pattern of thermal intensity; plain and simple. You are no longer simply looking at thermal, but visual aspects and this can be very confusing. It's not adding as much as distracting. Let's look at some examples:

    • Integrating visual aspects changes the intensity of colour observed in the thermal which can severely distort the information.
    • The “viewer” tends pays more attention to the highly detailed visual aspect of the image, rather than the most important aspect: thermal.
    • Limiting or resizing the thermal image over a visual background, disregards thermal content over visual. In many cases this means that similar or background components cannot be compared.
    • Verification of surface material and condition (ie emissivity) is not possible when visible image is obstructed by thermal.

    Here are some examples of misapplication of image blending in the above context...

    You can only look at 100% of anything at one time, and if you are observing visual aspects, then you are observing less than 100% of the thermal aspect!! Is that really what we want from a thermal survey?? Sure it’s impressive to those who don’t know better, and it looks more “visual”, but I’m not interested in visual. I want a detailed thermal pattern of my target that's why I'm paying the big bucks for thermal resolution.

    While I have seen some innovative uses of this function, it is more often abused than properly used.

    Does anyone actually remember why we include a visual image? In the old days it was because a thermal imager's resolution was so low, that if we didn't have a corresponding image we didn't know what we were actually looking at. Today's thermal cameras have sufficient resolution to clearly resemble the target.

    The purpose of a visual image in a thermal survey is not to tell us which component we are looking at, but to provide credibility to our measurement.

    The visual image allows the reader (end user) to identify/qualify a material’s apparent emissivity, surface condition (ie dirt, dust, oxidisation), sky conditions (ie overcast or clear sunny skies), whether it’s shaded etc … all the things that can adversely affect our measurement. It's there to make sure that the thermal data is credible. That when a thermographer says "here's a hot spot" we can discern that it's not a reflection or other error source. We don’t need a visual to make sure that a circuit breaker is actually a circuit breaker, or even to read it’s asset identification. This should be clearly identified by the thermographer’s documentation and recording of asset identification.

    Arguments of image identification only reflect a thermographers inability to meet the standard requirements. It is a thermographer responsibility to clearly identify a target, by it's asset identification number, circuit number, etc. It is not the responsibility of the end user to guess which target the thermographer is viewing.

    How cool is MSX?
    MSX is also a standard feature on all models. While there is no doubt it is graphically clever, my opinion is well established on features that provide a visual aberration… they distort and dilute the thermal information which is of primary importance.

    How many novices will inadvertently image through Perspex covers or glass believing they are seeing the components? This image below was taken directly through the Perspex, the visual component of the MSX feature displays the circuit breakers, yet the thermal camera will measure no further than the cover as infrared does not pass through solids (such as Perspex).

    Name:  FLIR-E8_msx-perspex_combo.jpg
Views: 181
Size:  20.7 KB

    With MSX you perceive you are seeing through a window.

    Name:  FLIR-E8_msx-window.jpg
Views: 179
Size:  24.4 KB

    Without MSX you can see you are measuring the window surface and the reflection of the thermographer.

    Name:  FLIR-E8_window.jpg
Views: 177
Size:  10.8 KB

    Name:  Flir-E8_msx_IR-vis.jpg
Views: 179
Size:  64.3 KB

    These features do however sell cameras and I respect the fact that the public want them (or they wouldn’t be there). MSX in particular gives the illusion of sharpening the image, so it does make the lower resolution models much more appealing to the eye. Anyone appraising a unit for purchase should turn off MSX to get a clear representation of the imager resolution.
    Last edited by Brenton; 03-27-2015 at 04:52 PM.
    Brenton Ward
    Level III Infraspection Institute Certified Thermographer
    Level II Airborne Ultrasound

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts