Imaging Subsurface Civil Infrastructure with Smart Seismic Network

The ability to use networked seismic instruments to image subsurface civil infrastructure and activities in data-limited extreme environments is crucial for many civil and security applications. This paper presents a non-invasive smart seismic network for imaging subsurface structures using ambient noise only. An innovative in-network spatial auto-autocorrelation method is designed to image different layers of underground infrastructures at different frequency bands and finally result in a 3D image. The proposed approach is general and can characterize the near-surface sediment and the infrastructure simultaneously. The experiments demonstrate that underground utility lines affecting sediment can be imaged, and the potential of using the method for abnormal activities like leakages. An exhaustive evaluation regarding bandwidth utilization, communication cost, and system resilience were conducted to highlight the benefits of the proposed approach.