I believe that you can do it, but the question is how large the aperture of your lens is. Aperture is the focal length divided by the f/no. So if you have an f=125 mm lens (real, not 35mm equivalent) and it opens up to f/2.8, your aperture is roughly 45mm.
This excellent answer mentions that the visual brightness of satellites in GEO is in the ballpark of +11 to +14 magnitude nominally, but can brighten considerably when the geometry is right, for example a solar panel or a metal or mirror surface. (See What are these very large, square panels on Inmarsat 5? and What is the function of this array of what looks like mirrors on TESS? for more about "mirror surfaces")
The answer links to satobj.org's Observing Geostationary Satellites
which says in part:
Unlike objects in low Earth orbit, geostationary satellites are visible throughout every night of the year, only entering the Earth's shadow for up to 70 minutes per day, around a couple of weeks either side of each equinox. During the same period the satellite tends to brighten over several days, twice a year, when the satellites orientation favors the 'beaming' of the Sun in the direction of the observer.
Typically the satellite will be in the mag. +11 to +14 range (or dimmer), but brightening by several magnitudes when the geometry is favourable (around mag. +5 to +6 is not untypical). One satellite is reported to have briefly been visible to the naked eye at mag. +3 !
So definitely do not point your camera directly away from the Sun (toward the anti-solar point) because any satellites in that direction would be in eclipse!
You will not resolve these objects, they will be points, like stars. For unresolved points, unlike extended objects, magnification doesn't affect brightness. So all you want is the largest possible aperture.
Read about astrophotography with SLRs to find out how to make sure you can focus your camera at infinity. Maybe a nearby star is the best, but there could be other techniques.
If you had a 40mm aperture pointed at the Sun (don't do that!) at ~ 1000 W/m^2, that would be of the order of 1 Watt at -27 magnitude. If you are looking at a reflected solar spectrum but at +10 magnitude, or 37 magnitudes dimmer, that would be $2.5 10^{-37/5}$ or about 1.6E-15 Watts. Like I did here, estimate 2.5 eV per photon and get about 6E-16 coulombs of e- per second in a bare silicon sensor, or about 100 e- per second.
Let's drop that a factor of 3 because your color CCD sensor has a subtractive bayer filter which rejects 2/3 of the light reaching each pixel.
Expose for 1 minute and you've got 2000 e-. That should be plenty if you are in focus and it's not smeared around too many other pixels.
The problem is your sky brightness, and unlike star brightness, that does scale with magnification. Higher mag means lower sky brightness.
So do some tests on your particular sky brightness.
Some information about measuring/classifying sky brightness can be found in answers to:
You might be able to find out if other amateur astrophotographers have any luck with filters to block sky brightness produced by city lights. Every city is different and it changes over time, but high pressure mercury and sodium lights tend to dominate and there may be filters that can help reduce those wavelengths that are affordable.
If you would like to learn more about sky brightness produced by local city lights, and how big "local" is, see the answers to Why (actually) is the night sky so bright in the city? How far up is that happening?
See this excellent answer for a clue where to look, which is just a few degrees south of the celestial equator. Chennai at 13.5 N latitude is about 1500 km above the equator, so looking out 36,000 km, you should look at about -2.4 degrees declination (south of the celestial equator).
Try several exposures of very different exposure times (steps of a factor of 2 or more) from short to very long, and repeat them after you move the camera by about 1/10 of a field of view. That way any real points will move but your hot pixels (if any) will remain fixed.
If you have a fancy camera you can program it to take a sequence of images. Maybe one hundred exposures of 5 seconds each. If you have that option, then you can just get all your data and play with it later. You can use image registration software to stack the images, there are several astrophotography-optimized image registration options out there to help you stack your images to get the best sub-pixel registration.
Just fyi, this cool answer to Has any ground-based telescope taken a picture of a geostationary broadcast satellite? shows that really big telescopes can take resolved images of satellites in GEO, but that's using adaptive optics.
Videos in Are commercial communications satellites in GEO being constantly monitored by telescopes? give a little bit of an impression of what's going on. Of course these are shot through telescopes, not camera lenses
