Tools in Development

RRA for Wind Hazards: The RRA can readily accept databases for other hurricane basins beyond than the Hawaii grid it was founded upon, and the surrogate modeling approach lends itself conveniently to spatial representations of wind fields during hurricanes, which are currently being developed for the platform using the same parametric description used for the wave and surge models.


Infrastructure Digitizer: The RRA workflow’s next stage of development will allow generation of Site Report Cards for assessment of the risk to individual assets due to combined wind, wave and surge loading. This will allow the identification of specific vulnerabilities within individual structures; exactly the type of actionable information necessary to drive risk mitigation by stakeholders. As the first step in supporting high fidelity risk assessments of individual assets, an automated Infrastructure Digitizer has been developed for CyberEye with specific focus on one of the most prevalent building typologies: light-frame wood residential construction. This add-on widget, whose workflow is visualized in Figure 4, assumes no access to structural drawings or other details of the home that are not publicly available and instead extracts the exterior geometry from publically available image sources and then constructs a finite element model from this information using a subassembly approach. Ground truth validation studies have confirmed that exterior geometries extracted from Google Street Views using doors as scaling features were found to have mean errors less than 10%, while the subassembly approach replicated the stiffness of the primary lateral system within 5%.


Reconnaissance App: The RESTful API lends itself readily to direct interfacing for digital reconnaissance. An iPad application (App) was recently beta-tested on Typhoon Haiyan/Yolanda reconnaissance in the Philippines. This App allows users to develop customized damage report forms by selecting from the available fields. These forms can be used through this App to execute completely digital field surveys that are stored locally on the device and then batch uploaded to the CyberEye Data Warehouse via WiFi or cellular connection, eliminating the need for manual uploading. This will create a more efficient mechanism to execute field investigations and a seamless means to encourage the community to contribute to the Data Warehouse.


Automated Damage Assessor: The next enhancements to data discovery will enable automated processing of aerial imagery for quantification of damage. The Automated Damage Assessor add-on widget consists of three steps: registration of objects of interest (buildings) from pre-event imagery, correction of photometric and geometric differences between before and after image-pairs, and then quantification of differences between these image-pairs to classify damage based on quantitative measures. Several advancements have been introduced in this framework to address the challenges associated with this process. The framework was evaluated on an expansive database of post-hurricane rooftop images to reveal a 93% accuracy in object registration, 85% accuracy in building extraction, and damage detection with 80% accuracy for a 3-scale damage metric and 72% accuracy for a fine grained 4-scale damage metric.

Suggested Reading

Kareem, A., Thomas, J., and Boyer, K. (2013). "Towards a Robust Automated hurricane Damage Assessment from High Resolution Images." Proc. 12th Americas Conf. on Wind Eng., June 16-20, Seattle, WA.

LaBarge, J., and Kijewski-Correa, T. (2013). "Rapid Infrastructure Digitization Framework to Support High-Fidelity Hurricane Risk Assessment." Proc. 12th Americas Conf. on Wind Eng., June 16-20, Seattle, WA.

Thomas, J., Kareem, A., and Bowyer, K. (2011). "Towards a robust, automated hurricane damage assessment from high-resolution images." Proc. 13th Intl. Conf. on Wind Eng., July 10-15, Amsterdam, Netherlands.