Amir

A very interesting question, and as a fellow keen photographer who regularly uses true HDR imaging. You have to look at your imaging technique from the beginning and insure that you are obtaining the images in the correct image format according to what the end outcome needs to be.

Could you clarify your set up?

Amir

I submitted a full explanation for you via the website, but somehow me when I clicked submit, it did not upload my answer, so here is my full answer to you, which you may need to get back to me with any update you can provide about what are you trying to achieve.

Photographic Answer

The CMOS Image sensor needs to be the largest and most sensitive you can get.

What camera body and lens are you using?

Does your camera body use any electronic filtration in front of the sensor that you can turn off.

Are you saving your images in uncompressed RAW format rather than JPEG?

- this is because the image size and depth is different. You need to be aiming for a 16 bit image where you have access to approx 16 million colors rather than a JPEG image which is usually compressed with a bit depth of 8 and a color range of 0-256

HDR Photography

This was developed due to the rather limited dynamic range of the photographic process. The human eye can see down to a very low light level, but of course even the best professional cameras can only get somewhere near what we can see. Our brains also change the data so that we see detail in dark parts of a scene as well as detail and texture in high light intensity scenes.

To take successful HDR images you need to know what your camera is able to do. You need to control what you can, and let the camera collect the information you cannot. In other words you need to have a fixed aperture (with its depth of field and focus) and bracket your exposures on either side of the median point by changing the exposure time. If you change the aperture between images you are varying the depth of field which will muddy your images and can cause unnecessary artifacts for form. There are however many different ways to achieve this end, so my advice is a practical one from my own experience.

If you have a Canon camera you may be able to take just one image either side of the best exposure setting, and a Nikon can get 2 or 3 either side. However, the number of images you take is not important, just that you expose for both detail in the low lights and detail in the high lights.

Most importantly, even if you have image stabilization available on your system, the use of a sturdy tripod and a remote release will prevent movement between shots and therefore mis-registration artifacts.

Software

The difference between True High Dynamic Range and HDR photography in the general sense is that software like Photomatix fuses the images and then applies what is known as Tone Mapping. This gives pleasing artistic results, as a plain True HDR image usually is perceived on a regular computer monitor as a rather dark muddy image.

A true HDR image is best obtained using the latest version of Photoshop (CC 2015). It fuses multiple images at 16 bit depth and contains the data you may be looking for in your research project ( which you have not enlightened us upon).

Hope that this is helpful to you, and have success in your project

Chris

Dear Dr. Flowers,

Thank you so much for your explanatory answer and guidance. Now I know what are the affecting parameters.

To make it more clear, I'm imaging a tissue phantom using a Canon D450. Light source is a red LED. I use tripod, but at the moment not sure about the lens I used.

The aim is to measure/analyse the light remitted form tissue/phantom at different wavelengths. This is my very primitive setup but I need to decide (using this initial setup) of what type of camera do I need to order, though not sure how these experiments could be helpful in this regard.

At the end I need to be able to compare images obtained through experiments with my Monte Carlo simulations.

I used JPEG images for the analysis I mentioned above.

The camera's position, ISO and F number were fixed. The camera was set to fully manual mode. I changed shutter speed from 1/1000 (I guess) to 1/10 of second and took 6 images, which I then blended in Photomatix.

Photoshop CC 0214 is installed on my machine. Actually I used Photoshop (blindly) but I didn't obtain results as same as Photomatix.

By the way, Photomatix gave a final image of six images which I opened it's RGB values separately in MATLAB to make sure I don't have saturation, and to observe a higher dynamic range which I couldn't see that!

Thanks very much again and looking for your further valuable advices.

Best regards,

Amir

Interesting. So you are looking at a form of near infra red imaging (NIR).

You need to decide whether your light source is in front of the tissue, or is used behind the tissue to analyze how the tissue alters the light beam.

I would suggest the second. Otherwise you would have trouble distinguishing reflected light from the light source itself and would need to shield the light source from the camera sensor.

When it comes to analysis, you still need the RAW data rather than a JPEG. If you save the data via MatLab, you could probably find a FREE Plug-In for MatLab that could accept a connection with your camera. In fact, that is probably the best way to move forward, and the cheapest. If you find a plug in that support data from a recent digital camera, you can buy that camera body. You do not need to worry about HDR at all, as the RAW data can be analyzed to it's full extent, without any proprietary filters or image algorithms being applied to the image data.

Did you realize that Photoshop integrates with MatLab, and the Adobe Developer Forum may be a good way to research this further. You may also be able to find someone who can develop a Plug-In for you and become part of your research team.

Regards, 

Chris

Right, as I guess I had to use RAW data.

Thanks very much again.

Best regards,

Amir