On Sat, Jun 29, 2013 at 12:01 AM, Eric Walker <[email protected]> wrote: > On Fri, Jun 28, 2013 at 11:04 AM, Berke Durak <[email protected]> wrote: >> So they want a custom-made camera with raw output from the sensor!? > > I doubt it. The way I read this point was that they wanted the raw data > feed from the IR camera rather than the analyzed data which presumably was > the output of software on the laptop.
It's still very unreasonable. It would mean that all the hard, camera-specific calibration work Optris & their software does would have been thrown away, to be replaced by something Levi et al. would do at great expense for a result with much lower confidence. First of all, the experimenters would need access to that "raw" data. Unless the Optris software allows the user to save some sort of raw data they would have to talk to the camera themselves, possibly requiring significant software development. Custom software means higher potential for error -- and deception. Then there is a lot of work between displaying pretty images and getting proper measurements. Every camera has to be individually calibrated. You need to know the transmittance of the optics and the response of the sensor. You need to properly set the amplifier parameters; corrections for dark current have to be applied. Non-uniform pixel response and optical effects such as vignetting have to be dealt with. Then you need to get rid of any remaining non-linearities and hope the instrument doesn't drift too much from the calibration curves you obtained at great expense. By the way, thermal sensors also need precise temperature control and continuous calibration against their own temperature, so it's not just static calibration. Levi et al. would have been real fools to attempt anything like that. Skeptics would have a field day. Calling the use of a known, calibrated instrument "an unnecessary detour" and suggesting instead a do-it-yourself temperature retrieval is quite strange (to put it mildly.) Basically they're like someone who is complaining because an electronic weighing scale has been used to measure a vertical force; they would rather have the experimenter read the resistance of the strain gauge directly instead of relying on the manufacturer-certified output of the instrument. -- Berke Durak
