After creating a mspec file my mean check of the grey standards in my photo do not always match the lrvs I have set for each of them. What are some potential reasons as to why this is happening and how concerned should I be (i.e. how “off” of the set lrv is too “off” for a set of measurements in the color channels?)? We would like to have absolute results to compare to future studies if possible, but will settle for relative results within this one experiment if necessary.
Potential ideas for the discrepancies:
- grey standards are submerged in water and photographed in air
- clear tape is used to keep the grey standards in place and prevents water from disintegrating them (standards created using paint swatches)
- some way in which the program calculates the measurements and possible the number of standards (n = 5 – I know it’s not necessary to use that many, but we wanted to try a couple of exploratory analyses) I used to create the mspec file
- Note: I did play around with setting different grey standards, different numbers of grey standards, and estimating the black point to see if the measurements for each standard were closer to their true lrvs, but didn’t have much luck. I’m going to play around with a couple of other iterations of test photos, but suspect the results will be no different.
Any and all suggestions will be greatly appreciated. Thanks!
Hi Jessica,
Once converted to a normalised image, if you measure the grey standard it should match almost exactly (<0.001% difference due to rounding errors) the value you input. If you’re using >2 standards then it plots a linear regression between them and there can be larger errors. All errors at this stage will be caused by the initial measurement of the standard being off, or some inconsistencies in the photo with >1 standard (such as differences in lighting between standards, which will be very common).
But it sounds like your issues are being caused by non-diffuse standards. It’s normally critical that standards are very diffuse, which is often difficult underwater. Any reflective barrier will make the standard non-diffuse. It is possible to use non-diffuse standards if you have very consistent lighting angle, or perfectly diffuse light. But in this case you shoudl be aware that while your photos might have good relative consistency, they are unlikely to have absolute reflectance precision.
Underwater I’d probably recommend using highly diffuse waterproof paint, or sanded/sandblasted marine grade stainless steel. You can also purchase ceramic grey standards which I’ve never used, but could be ideal for underwater work.
Cheers,
Jolyon