Robust methods for RICH ring recognition and particle identification

Problems of direct processing of raw Ring Imaging Cherenkov (RICH) data for the high granularity detectors of the CERES type are discussed. The robust technique is applied to ring fitting. It is based on a least-square method with recalculated weights depending not only on the distance of pads from the fitted circle but also on the signal amplitudes in pads. The proposed method leads to reliable parameter reconstruction from raw data for a single ring, as well as for two overlapping rings. Computed ring parameters are then successfully used for the particle identification.