New Powerful Methods for Photometry of CCD Images in Crowded Fields

Abstract

Image subtraction is an interesting new alternative to the classical profile fitting method (DAOPHOT or DoPHOT) for finding variable stars and producing their light curves. In crowded fields this new method can lead to large improvements in the photometric accuracy. The method is based on finding the best kernel solution in order to match two images as closely as possible. This approach leads to simple mathematical equations. It is possible to find a general solution to these equations which require very reasonable computing times. It is also shown that even in the case of a spatially variable kernel an optimal solution can be found with minimum computing time. Constant flux scaling can be imposed in this case without changing the basis of the algorithm. The method is illustrated using a set of images of the central region of the globular cluster M5. Only 26 variables were found by processing this data set with DoPHOT, while 61 were found with image subtraction. A large photometric improvement was also found for the 26 variables in common. The maximum improvement achieved by using image subtraction was a factor of 20 with respect to DoPHOT. The accuracy achieved with image subtraction is comparable to what was achieved with HST in a small region around the M5 center. One consequence of this photometric improvement was the discovery of an RR Lyrae star pulsating in a nonradial mode in M5. Finally, it is concluded that image subtraction is a technique of choice when dealing with variability, and that it is important to use it when the field is crowded. It is also important to note that image subtraction may open new possibilities in the investigation of very crowded fields, even with relatively small telescope from the ground.