An experimental survey of the estimation of stray light levels and corrections has been carried out. The performance of methods requiring no special instrumentation has been evaluated and compared. A new method, as convenient as the currently used filters method and capable of yielding better results especially when large stray light corrections are required, has been established and tested.