A KML-BASED APPROACH FOR DISTRIBUTED COLLABORATIVE
INTERPRETATION OF REMOTE SENSING IMAGES IN THE GEO-BROWSER
Liang Huang *', Xinyan Zhu *, Wei Guo *, Longgang Xiang *, Xu Chen *, Yifang Mei *
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing,
Wuhan University, Wuhan 430079, China
"State Key Laboratory of Software Engineering, Wuhan University, Wuhan 430079, China
Commission IV, WG IV/5
KEY WORDS: Image Interpretation, Distributed Collaboration, Collaborative interpretation, KML, GeoGlobe
ABSTRACT:
Existing implementations of collaborative image interpretation have many limitations for very large satellite imageries, such as
inefficient browsing, slow transmission, etc. This article presents a KML-based approach to support distributed, real-time,
synchronous collaborative interpretation for remote sensing images in the geo-browser. As an OGC standard, KML (Keyhole
Markup Language) has the advantage of organizing various types of geospatial data (including image, annotation, geometry, etc.) in
the geo-browser. Existing KML elements can be used to describe simple interpretation results indicated by vector symbols. To
enlarge its application, this article expands KML elements to describe some complex image processing operations, including band
combination, grey transformation, geometric correction, etc. Improved KML is employed to describe and share interpretation
operations and results among interpreters. Further, this article develops some collaboration related services that are collaboration
launch service, perceiving service and communication service. The launch service creates a collaborative interpretation task and
provides a unified interface for all participants. The perceiving service supports interpreters to share collaboration awareness.
Communication service provides interpreters with written words communication. Finally, the GeoGlobe geo-browser (an extensible
and flexible geospatial platform developed in LIESMARS) is selected to perform experiments of collaborative image interpretation.
The geo-browser, which manage and visualize massive geospatial information, can provide distributed users with quick browsing
and transmission. Meanwhile in the geo-browser, GIS data (for example DEM, DTM, thematic map and etc.) can be integrated to
assist in improving accuracy of interpretation. Results show that the proposed method is available to support distributed
collaborative interpretation of remote sensing image.
1. INTRODUCTION while large-scale interpretation task goes. Since interpreters lack
of effective communication and resources sharing, the accuracy
As the development of sensor and earth observation of image interpretation is deeply influenced by personal
technologies, earth observation system has taken shape and experience and professional knowledge background.
perfected gradually. Multi-scale and multi-spectral aerospace Inconsistencies exist among interpretation results without
remote sensing platforms carry out consistent observations of ^ communication. Thus the transformation of interpretation
earth. These platforms have continually producing multi-
temporal observational data. Tremendous data enrichment, calls
for efficient information extraction technologies to discover
underlying useful knowledge. Quantitative analysis and image
interpretation are two effective methods of extracting
information from remote sensing images (Richards and Jia,
2006). Image interpretation is a process of extracting target
information by analysing, comparing, reasoning various image
features comprehensively, based on imaging theories and
professional knowledge (e.g. Geology, Surveying and Mapping).
Image interpretation has experienced four stages of
development that are visual interpretation, interactive
interpretation, knowledge based expert interpretation and
automatic interpretation. During this process, image
interpretation methods have been more automatic and
intelligent. Practices of interpretation, however, have been
stand-alone operations. Multiple workers finish common
interpretation task in a serial way. As the improving of accuracy
and size of earth observation data, image interpretation task
becomes increasingly complex and difficult. Traditional mode
of serial working cannot satisfy the demands of efficiency,
*Corresponding author. Tel: +1 347 707 0410;
Email Address: plaquemine@àwhu.edu.cn (L. Huang)
practices is necessary. Collaborative interpretation is an
efficient way to improve efficiency and accuracy by cooperative
work. Multiple interpreters work synchronously or
asynchronously in a distributed environment can reduce
working time. Fully sharing of ideas and resources helps to
improve the accuracy of interpretation results. Computer
Supported Cooperative Work (CSCW) is one of the most
common technologies to implement cooperative work. With
CSCW, geographically dispersed researchers can complete a
task cooperatively. In the area of medicine, some scholars
introduce CSCW to support collaborative interpretation of
different medical image data for teleconsultation and analysis
(Andersen et al., 2011; Hu et al., 2007; Zhang et al., 2000;
Zhang et al., 2004). (Carlos et al., 2000) integrates CSCW and
image segmentation module to design a collaborative
environment called diSNet for volume data analysis. (Lim et al.,
2001) combines Java applets and CSCW to provide a web-
based collaborative system for medical image analysis and
diagnosis. (Huang et al., 2009) uses web services approach to
build a medical image synchronous collaborative analysis
system, which enables distributed physicians doing
207
