Combining DIVA images: detecting visual binaries with large magnitudedifferences

P. Nurmi

During the observational phase of the DIVA satellite several images are obtained from the same object at different epochs and transit angles. Hence, it is scientifically important to combine the images to produce one "super'' image that contains all the data. This allows us to find much fainter stars than those that can be detected in single images. Most of the emerging objects are physical secondary stars in binaries, but also some background objects will be found. The combination procedure also increases the image resolution, especially in the cross-scan direction. The combination procedure is somewhat more complicated than those used in traditional image combining problems. This is mainly due to two main differences. Firstly, DIVA pixels are binned 1 x 4 pixels and secondly, transit angles, i.e. image rotations, can have all the values between0 and 360degrees, which are generally not uniformly distributed. Also, the DIVA field is not a square, but it is elongated in the cross-scan direction. The developed image combination procedure uses a drizzling method as its basic idea. The drizzling method is used in many image combining problems that need to be fast and handle undersampled shifted and rotated images. We show some preliminary results of the image combination and illustrate the basic reduction procedure that could be used to detect faint companions around primary stars. By estimating the efficiency of the current image detection method we conclude that at distances of a few arc-seconds the drizzling method can reveal very faint objects and is especially efficient, but at sub-pixel level, the detection of faint companions becomes more difficult but not impossible. However, even in the worst possible scenario the combining procedure can provide a considerable amount of new binary detections.

Manuscript: jad9_8e.ps