Distributed Fiber Optic Sensing (DFOS): Principles and Applications

What Is Distributed Fiber Optic Sensing?

Distributed fiber optic sensing (DFOS) is a family of technologies that use optical fiber — the same glass strands that carry internet traffic — as a continuous sensing element. Unlike conventional sensors that measure at discrete points, DFOS transforms every meter of fiber into an independent sensor, enabling continuous monitoring over distances of 1 to 100+ kilometers from a single instrument.

The underlying physics is elegant: when coherent laser light propagates through optical fiber, it interacts with the glass at a molecular level. Tiny imperfections and density fluctuations in the glass scatter a fraction of the light back toward the source. By analyzing this backscattered light, a DFOS interrogator can detect, locate, and characterize disturbances anywhere along the fiber — with meter-level spatial resolution and millisecond temporal response.

The Physics of Rayleigh Backscatter

Three types of scattering occur when light travels through optical fiber: Rayleigh, Brillouin, and Raman. Each carries different information:

• Rayleigh scattering — caused by sub-wavelength density fluctuations frozen into the glass during manufacturing. Rayleigh backscatter is sensitive to strain and vibration, making it the basis for acoustic and intrusion detection. It provides the highest spatial resolution (down to ~1 meter) and fastest response time (kHz sampling rates).
• Brillouin scattering — caused by acoustic phonons (sound waves) in the glass. The Brillouin frequency shift is proportional to temperature and strain, enabling distributed temperature and strain measurement over long distances. Typical spatial resolution is 0.5–2 meters.
• Raman scattering — caused by molecular vibrations. The ratio of Stokes to anti-Stokes Raman components is temperature-dependent, providing distributed temperature sensing (DTS). Spatial resolution is typically 1–2 meters.

For security applications, Rayleigh-based sensing — specifically phase-sensitive optical time-domain reflectometry (φ-OTDR) — is the dominant technology. It detects the tiny phase changes caused by vibration, pressure, or acoustic energy impinging on the fiber.

Phase-Sensitive OTDR: How It Works

In a φ-OTDR system, a highly coherent laser emits short pulses into the fiber. Each pulse generates a backscatter trace — a one-dimensional signal representing the light reflected from every point along the fiber. Because the laser is coherent, the backscatter trace contains phase information: the light from adjacent scattering centers interferes constructively or destructively, producing a pattern that is unique and stable for an undisturbed fiber.

When something disturbs the fiber — a person climbing a fence, a vehicle driving over a buried cable, a jackhammer breaking concrete near an underground conduit — the local strain changes. This shifts the phase of the backscattered light from that section of fiber, altering the interference pattern. The interrogator detects this change by comparing successive backscatter traces at high speed (typically 1–10 kHz).

The result is a continuous, real-time measurement of vibration and acoustic energy along the entire fiber length, with spatial resolution of 1–10 meters depending on the system configuration.

From Raw Data to Actionable Intelligence

A φ-OTDR interrogator monitoring 40 km of fiber at 5 kHz generates enormous volumes of data — roughly 200 million phase measurements per second. Converting this raw data into actionable security alerts requires sophisticated signal processing:

• Noise suppression: Environmental noise (wind, rain, traffic, temperature drift) must be filtered without suppressing genuine intrusion signals. Adaptive filtering techniques — conceptually similar to the adaptive photoreceptors used in bio-inspired compound-eye research — continuously adjust to local noise conditions along the fiber.
• Event detection: Signal energy in specific frequency bands is monitored against adaptive thresholds. When energy exceeds the threshold, an event is declared and a time-frequency signature is extracted.
• Classification: Machine learning models (typically convolutional neural networks or random forests) classify the event signature as one of several categories: climbing, cutting, digging, walking, vehicle, animal, wind, rain, or machinery. Modern systems achieve classification accuracy above 95% with false alarm rates below 1 per km per day.
• Localization: The position of the event along the fiber is determined from the backscatter trace, typically to within ±5 meters. Multiple simultaneous events at different locations are resolved independently.

Security Applications

DFOS-based perimeter intrusion detection systems (PIDS) are deployed worldwide at the most demanding security sites:

• Critical infrastructure: Power plants, substations, water treatment facilities, and telecommunications hubs
• Transportation: Airports, rail corridors, pipelines, and port facilities
• Government and military: Military bases, government buildings, prisons, and border installations
• Commercial: Data centers, logistics hubs, pharmaceutical campuses, and high-value storage facilities

The technology's unique advantages — passive sensing (no electronics at the perimeter), immunity to electromagnetic interference, covert deployment capability, and massive scale (one interrogator covers what would require hundreds of conventional sensors) — make it the technology of choice for long-perimeter, high-security applications.

Beyond Perimeter Security

While security is the most visible application, DFOS technology is also used for structural health monitoring (bridges, tunnels, dams), pipeline leak and intrusion detection, railway monitoring, and geophysical sensing. The underlying physics is identical — only the signal processing and classification models change to match the application domain.

At Curvace, we focus on the security applications of DFOS, bringing deep understanding of the underlying optical physics to the design and deployment of systems that protect what matters most.