Feature-Based Simplification of Process Plant Models over Network

International audienceVisualization of 3D process plant models is an essential feature to support design review in distributed design environment. The preprocessing time, frame rate and visual fidelity play the same import role in improving the effectiveness and efficiency of the entire design review process. General model simplification algorithms, such as levels of detail (LODs) and hierarchical levels of detail (HLODs), always need unbearable time for preprocessing large-scale models and may distort design feature and make them unrecognizable to the reviewer. To efficiently transmit models, various compression technologies can be applied to shrink the data size. However, transmitting large-scale models over network is still a bottleneck of rendering performance. In this paper, we presented a new, faster feature-based model simplification algorithm to simplify process plant models over network. We first get the model’s geometric parameters and topology information from the server before visualization. Then we compute LODs and HLODs according to the model’s geometric parameters and composing feature on the client. No triangles are transmitted during visualization. We demonstrate its performance on complex process plant models composed of tens of millions of triangles. Results show that our approach is able to shorten the preprocessing time greatly. And it can achieve considerable speedups in frame rate with little loss in image quality