1201
SEAMLESS MAPPING SYSTEM FOR HIGH RESOLUTION IMAGERY
Haibin Ai a b *, Jianqing Zhang 3 , Yansong Duan b
‘‘School of Remote Sensing and Information Engineering, Wuhan University, 430074 Wuhan, Hubei, China
b Supresoft Incorporation, 430079 Wuhan, Hubei, China -
ahb32@163.com
Commission WgS-PS, Working Group IV/9
KEY WORDS: Digital Photogrammetry, High Resolution Image, Distributed Computing, Topographic Mapping, Interoperability,
Mapping System
ABSTRACT:
Digital Photogrammetry is the quality leap in the means, ability, efficiency and agility of spatial information acquirement. Digital
Photogrammetric Workstation (DPW) is the mainstream product of Digital Photogrammetry System. Nowadays DPW plays
important roles in acquiring spatial data. With the emergence of high resolution sensor, such as 1KONOS, QuickBird, SPOT5 and
ADS40, the traditional DPW doesn’t adapt current situation. Innovative digital photogrammetry system should be developed to meet
the network turn and industrialize production nowadays. At present, cartographic map collection is completed by means of manual
measurement or human-computer interactive half-automatic abstract. Manual measurement covers approximate 80 percents of the
cartographic map collection procedure. To efficiently mapping from high resolution images, a seamless mapping system based on
network is proposed.
1. INTRODUCTION
Digital photogrammetry automatically obtains objects’ 3D
coordinates by identifying the correspondences from two or
more digitized images or digital images. Comparing to Analogy
Photogrammetry and Analytical Photogrammetry, Digital
Photogrammetry is the quality leap in the means, ability,
efficiency and agility of spatial information acquirement.
Digital Photogrammetric Workstation (DPW) is the mainstream
product of Digital Photogrammetry System. Nowadays DPW
plays important roles in acquiring spatial data.
Cartographic Map, Digital Terrain Model (DTM) and Digital
Orthorectification Map (DOM) are important digital data
sources for Spatial Information Systems. DOM is automatically
produced from Digital Elevation Model (DEM) and original
images by sensor’s geometrical model. With the advance of
automatic image matching technique, DTM is also
automatically obtained from stereo image pairs. At present it is
not impossible that cartographic map is automatically collected
from images despite the advance of object abstract technique.
Cartographic map collection is completed by means of manual
measurement or human-computer interactive half-automatic
abstract. Manual measurement covers approximate 80 percents
of the cartographic map collection procedure, so it is difficult to
improve efficiency of the mapping.
With the inventing of modem satellite and airborne high
resolution sensors such as IKONOS, QuickBird, SPOT5 and
ADS40, high resolution images are larger than aerial images
and provide a challenge for algorithmic redesign, and this opens
up the possibility of reconsidering and improving many
photogrammetric processing components. In recent years,
considerable research efforts have been devoted to the efficient
utilization of these images. Research has covered, for example,
sensor modeling and image orientation, automatic digital terrain
model (DTM) and digital surface model (DSM) generation and
feature extraction [Li Zhang and Armin Gruen, 2006] .
Many photogrammetric processing approaches and workflows
have been proposed to resolve the challenges brought forward
by high resolution sensors’ images. Among these approaches
and strategies distributed and parallel processing is a very
important one. And a few distributed and parallel processing
photogrammetric systems are nowadays commercially available
in the market. For example, PixelFactory' developed by
Infoterra-Global company of France and PixelPipe 1 innovated
by Intergraph company of USA are outstanding distributed and
parallel processing systems. These innovated systems are
applied to practical photogrammetric produce and improve
produce efficiency of high resolution images. These systems
emphasize the automatic processing for high resolution images
such as DTM/DSM generation and orthorectification.
Unfortunately, there are no corresponding packages to collect
cartographic map from stereo pairs in these systems.
Cartographic map is a very important photogrammetric product.
Unreservedly speaking, these systems are immature systems
because of lack of cartographic map collection toolkits.
Seamless mapping system proposed in this paper is a sub
system of DPGrid which is a distributed and parallel
photogrammetric processing system innovated in Wuhan
University. As a new generation digital photogrammetry system,
DPGrid consists of an automatic processing sub-system and a
human-computer interactive processing sub-system. The
objectives of automatic processing sub-system are similar to
PixelFactory 1 and PixelPipe 1 which are used to automatic
processing. Seamless mapping system is the human-computer
interactive mapping system which is designed to collect
cartographic map from stereo pairs. There are a few
* Email: ahb32@163.com Tel: 13607138724