Statistical recognition methods and robust estimation of track parameters

Mathematical foundation of particle track recognition algorithms in track chamber data analysis is considered. The recognition is interpreted as problem for some nonhomogeneous population sample estimation. Such an approach makes it possible to apply the robust estimation methods. The development of these methods that allows one to improve essentially the accuracy and the speed of the tracking procedure is suggested. This was confirmed by solving model problems. Filtering algorithms based on the suggested methods and their applications to process real data scanned from streamer chamber pictures are described