The concept of reliability applied to co-registration of multiple satellite time-series of images

This paper wants to explore the application of the theory of reliability, which has become largely popular in the analysis of geodetic and photogrammetric networks, to the multiple co-registration of satellite time-series of images through block adjustment. Indeed the current and future availability of long series of satellite imagery can be exploited for the purpose of change detection and soil monitoring along time. Due to the large overlap between images, the multiple co-registration is a feasible task that may result in an increased reliability of the observations. On the other hand, to deal with such big data sets, automatic procedures have been developed for co-registration of satellite time-series. Due to the presence of residual errors in such procedures, the error analysis during the final block adjustment plays a key-role. Consequently, the background on the theory of reliability has been introduced here, and three simulated examples have been created in order to check the efficiency of the reliability analysis when dealing with time-series of medium resolution satellite imagery.