Uncertainty in mapping of typical wetland plants using remote sensing images

It is helpful for wetland ecological environment monitoring work to extract the typical wetland plants growth and distribution information exactly. There is much uncertainty in wetland plants information extraction using remote sensing images due to the complexity of wetland plants and remote sensing data. Recently, remote sensing data have become one of the most important data sources in the Ecology and the Environment research area and therefore researches on uncertainty in wetland plants information extraction from remote sensing data are essential. This study presents a case study conducted at the core area of Honghe National Nature Reserve in the Sanjiang Plain, Northeast China. The SPOT 5 image was selected to produce the Carex lasiocarpa map by using object-oriented approach. And the map was validated by using the method of confusion matrix based on verification points composed of ground truth points obtained by a Differential Global Positioning System with a positional accuracy of 0.02 m and random points generated by a GIS (Geographic Information System) method. Also, by using the same approach, the Carex lasiocarpa map was produced by the airship image whose acquisition date was in the same growing season with SPOT 5 image´ acquisition date. We analyzed the uncertainty in the Carex lasiocarpa information extraction from SPOT 5 image based on the Carex lasiocarpa map derived from the airship image from two perspectives of patch number and special patch attribute and achieved the following conclusions: 1) The Carex lasiocarpa map derived from the SPOT 5 image represented space distribution pattern of Carex lasiocarpa in the study area and meanwhile some mission extraction and wrong extraction existed; 2) Only those Carex lasiocarpa patches whose area was larger than 160.69 m² could be extracted from the SPOT 5 image; 3) some errors existed in the position, patch outlines shape and patch area on the Carex lasiocarpa information derived from the S- OT 5 image.