Techniques to retrieve information about orbits of dynamical systems l
ost in the process of making measurements due to discretization, sampl
ing, or noise are developed. A general method of reprocessing the expe
rimentally obtained data to effectively increase its information conte
nt (the accuracy of measurements) is proposed. Potential applications
include many areas of applied mathematics such as inverse problems, ti
me series analysis, ill posed problems, etc. The method can also be us
ed to help construct new highly accurate measuring devices.