Calibration of illuminance using canoncap and HDRcapOSX vs Photosphere

Hi Brent,

First off, how do you use your illuminance meter to calibrate your Photosphere image? Since Photosphere needs a *luminance* value to calibrate from, you'll need to convert from illuminance to luminance. For this to work, you need a diffuse, gray surface of known reflectance, such as an 18% gray card. Measuring the illuminance at the gray card (and in the same orientation as the gray card), you can then compute luminance from:

  Luminance = Illuminance * 0.18 / 3.1416

This of course assumes that the illumination doesn't change between the time you measure it and the time you capture the scene.

Once you are doing this correctly, you can apply a scaling factor equal to the average ratio in the image from Photosphere and the identical scene capture from canoncap. In fact, you don't really need Photosphere at that point. You just need the ratio of your actual luminance to that reported by canoncap.

Once you have this ratio, you can use the pcomb program to adjust the exposure like so:

  pcomb -o -s [ratio] input.hdr > output.hdr

There are other, quicker and more clever ways to do it using getinfo and sed, but the above is the simplest.

Cheers,
-Greg

···

From: Brent Huchuk <[email protected]>
Date: June 26, 2013 6:15:34 PM PDT

Hi everyone -

I have just started using and testing HDR images for glare analysis as part of my research. The question I face is that when I manually generate an image using Photosphere I am able to calibrate the image using an illuminance meter which, when false colours are applied, generates one range of values. If I duplicate this scene using canoncap an HDR image is output but one that gives very different (like 3 times smaller) upper luminance values. Is there a part to the method I have not accomplished and/or is there a way to calibrate the process in order to get these illuminance values to match the calibrated process?

I am hoping this to be the case so I can set up a time-lapse situation like others have successfully done and process the images using evalglare.

Thanks for any assistance.

Cheers,

Brent

Brent Huchuk
Carleton University
MASc Candidate

Hi Brent,

HDR glare analysis is VERY sensitiv to the accuracy with which the
highlights in a scene are captured. You MUST make sure the dynamic
range of your sequence is sufficient. WebHDRtools come with a program
called jpgheatmap, which is a Perl script. When you run it against
your sequence, you should not have any red dots on the heatmap. Blue
ones are all right.
http://www.jaloxa.eu/resources/hdr/webhdrtools/index.shtml
You can try it on-line:
http://www.jaloxa.eu/webhdr/roll-your-own.shtml

Cheers

Axel

···

On 27 June 2013 05:07, Gregory J. Ward <[email protected]> wrote:

Hi Brent,

First off, how do you use your illuminance meter to calibrate your Photosphere image? Since Photosphere needs a *luminance* value to calibrate from, you'll need to convert from illuminance to luminance. For this to work, you need a diffuse, gray surface of known reflectance, such as an 18% gray card. Measuring the illuminance at the gray card (and in the same orientation as the gray card), you can then compute luminance from:

        Luminance = Illuminance * 0.18 / 3.1416

This of course assumes that the illumination doesn't change between the time you measure it and the time you capture the scene.

Once you are doing this correctly, you can apply a scaling factor equal to the average ratio in the image from Photosphere and the identical scene capture from canoncap. In fact, you don't really need Photosphere at that point. You just need the ratio of your actual luminance to that reported by canoncap.

Once you have this ratio, you can use the pcomb program to adjust the exposure like so:

        pcomb -o -s [ratio] input.hdr > output.hdr

There are other, quicker and more clever ways to do it using getinfo and sed, but the above is the simplest.

Cheers,
-Greg

From: Brent Huchuk <[email protected]>
Date: June 26, 2013 6:15:34 PM PDT

Hi everyone -

I have just started using and testing HDR images for glare analysis as part of my research. The question I face is that when I manually generate an image using Photosphere I am able to calibrate the image using an illuminance meter which, when false colours are applied, generates one range of values. If I duplicate this scene using canoncap an HDR image is output but one that gives very different (like 3 times smaller) upper luminance values. Is there a part to the method I have not accomplished and/or is there a way to calibrate the process in order to get these illuminance values to match the calibrated process?

I am hoping this to be the case so I can set up a time-lapse situation like others have successfully done and process the images using evalglare.

Thanks for any assistance.

Cheers,

Brent

Brent Huchuk
Carleton University
MASc Candidate

_______________________________________________
HDRI mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/hdri