Uncertainty analysis of object location in multi-source remote sensing imagery classification

Abstract

In object-oriented multi-source remote sensing imagery classification, it is an essential prerequisite to locate objects on different images. The spatial uncertainty of located object boundaries is unavoidable and may have a significant impact on the subsequent object feature calculation and classification. To seek the proper object location scheme, the image resampling and the transfer of object boundaries are studied by uncertainty impact analysis. Results indicate when images are resampled to high spatial resolution, the object statistical features and classification accuracy are little affected by the object boundary uncertainty; transfer of raster or vector object boundaries are both adoptable. Whereas when images are geo-registered without changing spatial resolution boundary uncertainty has a significant influence on the statistical value of texture feature and tends to induce the instability of classification results, the object locational uncertainty cannot be disregarded unless it is controlled in a certain limited range.