Quantifying light intensity from an image
Posted: 2014-09-24T03:02:01-07:00
Hi,
I would like to get a measure of the absolute light intensity striking a detector, but I do not have access to the voltage output of that detector, only the output in the form of an image. I can read this output as the value of a pixel as a number between 1 and 2^16 (i.e. 65536). There are two settings that I can adjust when acquiring the image - brightness and contrast. I am presuming that in order to create the image from the acquired light intensity, the output is a straight line, ie y = mx + c, where x is the voltage from the sensor, c is related to the brightness setting, m is related to the contrast setting and y is the number in the image for a particular pixel. I am thinking that if I put filters over the sensor so that I can cut the amount of light by one half or one quarter, I could determine m and c for a particular brightness and contrast setting of the detector and thus I could determine the light intensity (or at least a value proportional to the light intensity) for any future measurement using any pixel in the image.
I suspect things cannot be as simple as this, but has anyone tried to determine what the light intensity actually is directly from an image or series of images?
Thanks
Chris
I would like to get a measure of the absolute light intensity striking a detector, but I do not have access to the voltage output of that detector, only the output in the form of an image. I can read this output as the value of a pixel as a number between 1 and 2^16 (i.e. 65536). There are two settings that I can adjust when acquiring the image - brightness and contrast. I am presuming that in order to create the image from the acquired light intensity, the output is a straight line, ie y = mx + c, where x is the voltage from the sensor, c is related to the brightness setting, m is related to the contrast setting and y is the number in the image for a particular pixel. I am thinking that if I put filters over the sensor so that I can cut the amount of light by one half or one quarter, I could determine m and c for a particular brightness and contrast setting of the detector and thus I could determine the light intensity (or at least a value proportional to the light intensity) for any future measurement using any pixel in the image.
I suspect things cannot be as simple as this, but has anyone tried to determine what the light intensity actually is directly from an image or series of images?
Thanks
Chris