Using ROTSE-III data, I identified a variable star. I hope to use SuperWASP data to generate a more accurate light curve for this star, but the SuperWASP data is heavily blended. I am trying to deblend the data using Sextractor, but the .fits files available online at https://wasp.cerit-sc.cz/search are in a table format that Sextractor cannot handle. Do the data exist in an image format? If so, where can I find these images? Is there another de-blending method that works well for SuperWASP data?
You deblend using the same way you do with other survey data. SuperWASP has photometric aperture of 55", so there are multiple stars affecting your lightcurve, more than ASAS-SN needs. You need to sum their magnitudes (from VizieR, DSS images on Aladin Lite help to identify them) and treat like a single star. Final value will be used to deblend SuperWASP magnitudes.
I have tried to phase the ASAS-SN data several times for this system, but I have never gotten a clear cut light curve like the one you uploaded. I went back a couple hundred days on https://asas-sn.osu.edu to find the data and then I only used data taken with a V filter. The gnuplot is for a phased period of 4.418844 (double the original as you suggested), but it differs greatly from your plot. What am I doing wrong here?
The period in the plot uploaded by Vitali is 4.3998 not 4.418844 d. so your period is not correct.
Also, you have to enter 2700 or so in the "Enter number of days to go back" field to get all the available data, not only a couple of hundred days.
I need an advice.
Magnitudes for what bandpass should I use to deblend SuperWASP data? (I would superimpose them with ASAS-SN, I'm aware that anyway, some additional empirical shift will be needed)
SuperWASP gathered data as CV, may I use V magnitudes for contaminating stars to deblend? Or GAIA G would be better?
Using V would be okay unless the star is too red. Remember that you are combining the data and then shifting it to the V zero point only for period determination purposes so the actual range will be determined from the V data alone. So any small difference in amplitude between datasets caused by a slightly different magnitude for the contaminating star won't affect the final numbers. Also, the impact of a small difference will only be important when blending is really strong.