3D Fire Front Reconstruction Based on Multi-View Observation and Complex Uncertainty Model.

This work presents a novel method of 3D reconstruction of the forest fire front in the tactical forest fire monitoring system based on uncertain multi-view observations. The proposed method uses the hierarchical virtual boxes structure instead of the discretized spatial model that allows us to exclude iterative "heavy" geo-localization and geo-rectification processes from the pre-processing stage. The fire front representation is based on the complex uncertainty model that allows us to smooth the effects of distortions, interferences, and noises. The complex uncertainty model is based on the simultaneous use of the fuzzy, rough, level soft sets as well as grey numbers. It takes into account the dynamics of the fire front spreading and provides a proper visual picture of the reconstructed 3D model of the forest fire. The proposed solutions provide an opportunity to speed up the fire front reconstruction process and improve the quality of visualization for the decision-maker allowing him to focus on a stable core of the process and reducing the impact of distortions and interferences. The developed method provides sufficient performance to fulfill the requirements of real-time forest fire response operations.