Semantic-Based Remote Sensing Images Intelligent Service on Grid Environment

In this work, we studied how to rapidly match remote sensing images by the semantic information of geographical objects in Grid architecture and how to slice, index, and assemble each tiled image in every grid node. We first designed the grid architecture of remote sensing images sharing and gave a new idea that searching corresponding images by semantic information of geographical objects. To each grid node, we put forward a new method to partition, index, organize and assemble remote sensing images by rectangle cell. Then, a prototype system was set up: five grid node machines and a grid server were employed to establish the remote sensing images service grid environment. Massive remote sensing images were treated with new methods above in grid node and vector data managed in grid server to supply semantic information. The tested results proved that goal remote sensing images were searched with higher speed and more accuracy in grid environment and each grid node had faster response compared with traditional remote sensing images sharing mode. And more intelligent services of remote sensing images were given for the multi-resolution, multi-temporal tiled images can be dynamically assembled and analyzed. Experiments proved it was a promising solution to provide remote sensing images services using Grid architecture in light of actual circumstances.