Accepted by MNRAS, 8 pages, 4 figuresInternational audienceThe numerical kernel approach to difference imaging has been implemented and applied to gravitational microlensing events observed by the PLANET collaboration. The effect of an error in the source-star coordinates is explored and a new algorithm is presented for determining the precise coordinates of the microlens in blended events, essential for accurate photometry of difference images. It is shown how the photometric reference flux need not be measured directly from the reference image but can be obtained from measurements of the difference images combined with the knowledge of the statistical flux uncertainties. The improved performance of the new algorithm, relative to ISIS2, is demonstrated

PLANET (the Probing Lensing Anomalies NETwork) is a worldwide collaboration of astronomers whose primary goal is to monitor microlensing events densely and precisely in order to detect and study anomalies that contain information about Galactic lenses and sources that would otherwise be unobtainable. The results of PLANET's highly successful first year of operation are presented here. Details of the observational setup, observing procedures, and data reduction procedures used to track the progress in real time at the three participating observing sites in 1995 are discussed. The ability to follow several events simultaneously with a median sampling interval of 1.6 hours and a photometric precision of better than 0.10mag even at I=19 has been clearly demonstrated. During PLANET's 1995 pilot campaign, 10 microlensing events were monitored; the binary nature of one of these, MACHO 95-BLG-12 was recognized by PLANET on the mountain. Another event, OGLE 95-BLG-04, displayed chromaticity that may betray the presence of blending with unresolved stars projected onto the same resolution element. Although lasting only about a month, the campaign may allow constraints to be placed on the number of planets with mass ratios to the parent star of 0.01 or greater.

We analyze PLANET photometric observations of the caustic-crossing binary-lens microlensing event, EROS BLG-2000-5, and find that modeling the observed light curve requires incorporation of the microlens parallax and the binary orbital motion. The projected Einstein radius (r E = 3.61 ± 0.11 AU) is derived from the measurement of the microlens parallax, and we are also able to infer the angular Einstein radius (θ E = 1.38 ± 0.12 mas) from the finite source effect on the light curve, combined with an estimate of the angular size of the source given by the source position in a color-magnitude diagram. The lens mass, M = 0.612 ± 0.057 M ⊙ , is found by combining these two quantities. This is the first time that parallax effects are detected for a caustic-crossing event and also the first time that the lens mass degeneracy has been completely broken through photometric monitoring alone. The combination ofr E and θ E also allows us to conclude that the lens lies in the near side of the disk, within 2.6 kpc of the Sun, while the radial velocity measurement indicates that the source is a Galactic bulge giant.

We present a method to analyze binary-lens microlensing light curves with one well-sampled fold caustic crossing. In general, the surface of χ 2 shows extremely complicated behavior over the 9-parameter space that characterizes binary lenses. This makes it difficult to systematically search the space and verify that a given local minimum is a global minimum. We show that for events with well-monitored caustics, the caustic-crossing region can be isolated from the rest of the light curve and easily fit to a 5-parameter function. Four of these caustic-crossing parameters can then be used to constrain the search in the larger 9-parameter space. This allows a systematic search for all solutions and thus identification of all local minima. We illustrate this technique using the PLANET data for MACHO 98-SMC-1, an excellent and publicly available caustic-crossing data set. We show that a very broad range