To date

Remote Sensing technologies have been applied to environmental management

monitoring and control. Remote Sensing technologies are also being applied to monitor land desertification

land use cover

environment pollution

etc. Due to limitations in instrument and processing technology in RS

system errors and measurement errors may corrupt the data. Errors may be associated with both attribute value and its location. After collecting spatiald ata

we then need to process

analyze and convert the data in order to make RS information understandable to users. When we produce the final RS products

new errors can be created as a result of spatial operations on maps or images that contain errors （spatial error propagation）. If we ignore these errors in the RS products

it can result in economic loss. Though

the uncertainty of remote sensing information has been a hotspot in the spatial information process and many research have provided us lots of valuable results

however very few of them is on the mechanism based. The existing methods applied to the uncertainty analysis of RS will lead to the blind spot where uncertainty could not be analyzed

and uncertainty in subsequent image product will be propagated

tracked and quantified inconveniently. In this paper

we propose a framework to deal with these problems.