However, since image processing was

However, since image processing was not the main focus of this study, a simplified manual method was used to show how an image processing approach can be included within the proposed approach. The snapshots retrieved from the video camera were processed manually by using a CAD software to determine how much of the available passage was blocked in each node following a damage. The output obtained from the video camera is a percentage (i.e., blockage ratio) showing the proportion of the unblocked areas retrieved from a snapshot (1) before any damage was simulated and (2) after some damage was exerted to the building components.Fig. 5 shows the unblocked areas that were detected for each node by using a snapshot of the hallway when the bookshelf that was located in Node 2 has fallen down. The example in Fig. 5 presents a local damage that only affects Node 2 and does not cause blockage in other nodes. However, when a damaged object is in the nearest node to the video camera (e.g., suspended ceiling in the third node) and is completely blocking its view, it is assumed that all three nodes are completely blocked since it is not possible to identify the amount of blockage at each node in such a case.