>Depth of field, the focal length of the objective and the working distance.
The cantilever oscillation amplitude is in the range of submicron or micron, so for the 8 mm focal distance lens, you can neglect it.
>Since the view followed by the dichroic mirror, it is difficult for me to understand, what depth of field should I use and how to calculate it in case of a dichroic mirror?
About dichroic mirror, there are several aspects. Dichroic means only spectral dependence of reflection/transmission ratio. So one spectral range, say laser one, will be mainly transmitted, while white light will be mainly reflected.
In the layout you presented, a dichroic mirror is unnecessary because of monochromatic laser light. It can be a conventional semitransparent mirror.
From the other side, there is no white light source, which will illuminate the sample, cantilever to correctly arrange the measurements. With such additional illumination application of a dichroic mirror is practical. It should decrease the influence of such illumination on the photodiodes signals.
>Further more a light source would be necessary to have a good view using the camera. Should I use a condenser to distribute the light evenly in the cantilever area ?
Yes, I mentioned such an illuminator above. Is the presented layout fixed, or do you have the possibility to modify it? It is not a strong requirement to obtain a homogeneous distribution of white illumination light. It can be used even image of the white light source (lamp or LED) on the sample, cantilever area.
>How can I make sure that footage taken by the camera is not white ?
I don't understand your question. If you use a correct dichroic mirror, the camera will image a white illuminated sample with a cantilever and a red laser spot on it.