Well, if your detection behaves ideally, then the number of detected photons, thus the intensity, should be proportional to the exposure time. However, in reality the cameras have some dark offset, which is independent of the exposure time (the background signal you get when there is no light going to the camera) . So the dependence of intensity on exposure time doesn't go to zero when extrapolated to zero exposure time. In your situation I would probably image a very simple and stable sample (such as an aqueous solution of a stable fluorophore, or maybe even better constant transmitted illumination) with different exposure times and see the dependence of intensity of the images on the exposure. That's the best course I imagine not to waste your data completely. Of course the real experiment should be still repeated with uniform exposure times in the end.

Well, if your detection behaves ideally, then the number of detected photons, thus the intensity, should be proportional to the exposure time. However, in reality the cameras have some dark offset, which is independent of the exposure time (the background signal you get when there is no light going to the camera) . So the dependence of intensity on exposure time doesn't go to zero when extrapolated to zero exposure time. In your situation I would probably image a very simple and stable sample (such as an aqueous solution of a stable fluorophore, or maybe even better constant transmitted illumination) with different exposure times and see the dependence of intensity of the images on the exposure. That's the best course I imagine not to waste your data completely. Of course the real experiment should be still repeated with uniform exposure times in the end.

Hello Aparna,

you could normalize the images, by dividing the pixel intensity on the exposure time as long as you are working in the linear range of the camera. Here is the description of the procedure:

1) you need to determine the dark pixel intensity, by acquiring 10 (or more) images for each exposure time (chose exposure time range, e.g. 10 ms, 50 ms, 100 ms, 200 ms, 500 ms, 1000 ms) without sample with camera shutter closed. Then calculate the mean intensity for each exposure time from the 10 images you acquired, you could use ImageJ or similar program for this purpose.

2) measure as it was suggested aqueous solution of fluorophore (you could use either different solution concentrations and use one and the same exposure time or use one concentration and measure with the range of exposure times). Make sure not to use saturated images. Again calculate mean pixel intensity for each set of images. Plot pixel intensity of lower exposure time vs pixel intensity of higher exposure time and check if you have clear linear relation by calculating correlation coefficient or determination coefficient. If you have liner relation, then you could proceed further with normalization.

3) You could use following ratio to normalized your images, by choosing one exp. time as reference and normalized against it: (I_exp1 – I_darkpxl)/t_exp1 = (I_exp2 – I_darkpxl)/t_exp2, where I_exp1 and I_exp2 are the experimentally measured pixel intensities, I_darkplx – mean dark pixel intensity, t_exp1 and t_exp2 are corresponding experimentally used exposure times.

Dear Radek,

Thank you for your suggestion, this is what I also thought at very beginning .

But there is so much fluctuations that it is bit tricky to correlate.

But thank you for your feedback.