SIMBAD references

We present an algorithm that can detect blends of bright stars with fainter, unassociated eclipsing binaries. Such systems contaminate searches for transiting planets, particularly in crowded fields where blends are common. Spectroscopic follow-up observations on large-aperture telescopes have been used to reject these blends, but the results are not always conclusive. Our approach exploits the fact that a blend with an eclipsing binary changes its shape during eclipse. We analyze original imaging data from the Optical Gravitational Lensing Experiment (OGLE), which were used to discover planet transit candidates. Adopting a technique developed in weak gravitational lensing to carefully correct for the point-spread function, which varies both with time and across the field, we demonstrate that ellipticities can be measured with great accuracy using an ensemble of images. Applied to OGLE-TR-3 and OGLE-TR-56, two transit systems, we show that both systems are blended with fainter stars, as are most other stars in the OGLE fields. Moreover, while we do not detect shape change when TR-56 undergoes transits, TR-3 exhibits a significant shape change during eclipses. We therefore conclude that TR-3 is indeed a blend with an eclipsing binary, as has been suggested from other lines of evidence. The probability that its shape change is caused by residual systematics is found to be less than 0.6%. Our technique incurs no follow-up cost and requires little human interaction. As such, it could become part of the data pipeline for any planetary transit search to minimize contamination by blends. We briefly discuss its relevance for the Kepler mission and for binary star detection.