This paper presents a GA(Genetic algorithms)-based band selection procedure for producing a Color composite Ratioing image with Maximum Entropy (CRME image) using hyperspectral data (e.g., hyperion data). The more number of the hyperspectral bands, the more dicult the ecient and eective processing for producing color composite images. In producing the CRME image, one of the requirements is to maximize the amount of information in the images (i.e., image entropy). Through the GA operations, the increasing of image entropy of CRME image (i.e., tness value) was conrmed, which corroborates the GA operations can be applied for band selection in producing the CRME image. Toward the end of run in the GA operations, three ratioing images in case of maximizing the entropy of CRME image are selected and based on the spectrum reection characteristic, those are assigned to red-, blue-, and green-plane, respectively. Compared with the true, natural and false color composite images, we conclude the produced CRME image is useful for interpretation of image features in terms of the amount of information in the image.

View full abstract

Onboard SPOT 5, the HRS instrument systematically collects 600 km × 120 km stereopairs around the Globe since 2002. From this time, SPOT5 stereoscopic imagery becomes one of main satellite data sources for accurate DTED level 2 DEM extraction.
Spot Image and French National Cartographic Institute (IGN) decided in 2002 to build a worldwide accurate database called Reference3D^TM using HRS data. Huge eorts have been made to standardize the process in order to oer aordable prices. Until 2008, the targeted accuracies were 16 mCE90 for horizontal accuracy, and 10 mLE90 for vertical accuracy. These accuracy requirements were achieved without any control points nor map support.
The introduction of the paper briey reminds the most signicant assessments performed by major players within the geospatial community, and more specically the one by ImageONE Co., Ltd. (Tokyo) on two Reference3D geocells over the Northern coast of Japan, Hokkaido province. This work was published in 2008 during the last Beijing ISPRS congress.
In 2009, it was decided to introduce reliable GCPs within the Reference3D production process, to increase the horizontal accuracy down to 6-10 mCE90. In addition, two new layers were added to the product, which provide the user with i) the horizontal accuracy for every single pixel of the Reference3D orthoimage and ii) the vertical accuracy for every single elevation value within the Reference3D DEM.
The paper shows how the extensive use of ICESat data brings valuable information regarding the eective vertical accuracy, and how ICESat data allows to quantify the elevation accuracy of a dataset.
In conclusion, we present the road map for the update of the whole Reference3D database, which currently spreads over more than 45 millions of sq. km. (being more than 4,200 1° by 1° geocells), already funded and started up to 2014, towards 80 Mkm² of Reference3D products.

View full abstract