Acoustic Imaging

What Is Acoustic Imaging?

Acoustic imaging is the camera-based form of sound inspection: an array of microphones records the sound field in front of it, software calculates where each frequency component originates, and the result is a heat map of noise sources laid over a normal photograph. The technique extends human hearing into the ultrasound range, above 20 kHz, where compressed air leaks, electrical discharge, and early-stage mechanical friction produce their strongest signals.

The practical value comes from localization. A single handheld ultrasound probe tells you that a leak exists somewhere nearby; an acoustic imaging camera shows the leak as a colored spot on the exact pipe fitting, insulator, or bearing housing it comes from. That difference is what turned ultrasound inspection from a specialist skill into a survey tool a maintenance team can walk a plant with.

How Acoustic Imaging Cameras Work

The Microphone Array

An acoustic camera is a flat panel or fan-shaped device carrying an array of microphones, typically dozens of MEMS (microelectromechanical systems) units per device, with many commercial arrays carrying 64 or more. MEMS microphones are tiny, cheap, and consistent, which is what makes it possible to fit so many onto a panel the size of a tablet and sample all of them in sync. Every microphone records the same scene from a slightly different position, and that spacing is what gives the software its ability to work out direction.

Beamforming: Steering a Virtual Microphone

Beamforming is the core algorithm. Software applies tiny time delays to each microphone's signal so that sounds arriving from one specific direction add together constructively while sounds from every other direction cancel out. The result behaves like a highly directional virtual microphone that can be pointed anywhere in the camera's field of view. The software sweeps this virtual microphone across thousands of points in the scene and computes a sound intensity value for each one, producing the acoustic heat map.

Frequency Filtering

Raw sound at a plant is a mix of process noise, machine hum, and the target signal. Before beamforming, the operator selects a frequency band, and the camera analyzes only that slice. Leak hiss, electrical discharge, and bearing friction each concentrate in different parts of the spectrum, so band selection is both a noise filter and a way to tell signal types apart. Most industrial cameras let you switch bands on the fly during a scan.

The Sound Map Overlay

The heat map is drawn over a live photo or video feed from a regular optical camera, usually in a red-to-blue color scale where red marks the loudest source. The operator sees, in real time, a colored blob sitting on the specific valve, coupling, or bushing where the noise originates. Modern cameras can record video clips of the overlay, tag the asset, and export the image for the work order.

What Acoustic Imaging Detects: Frequency Bands

The ultrasound band carries most of the useful industrial signal, but different fault types sit in different sub-bands. Operators pick the band that matches what they are hunting.

Each source type also has a recognizable time pattern on top of its band. A gas leak produces a steady broadband hiss. A steam trap cycling open and closed pulses. Corona on a high-voltage insulator produces a continuous crackle concentrated above 40 kHz, while arcing throws short, sharp spikes. Skilled surveyors read both the frequency and the pattern before classifying a finding.

Industrial Applications of Acoustic Imaging

Compressed Air Leak Surveys

Compressed air is the most common starting point, and for good reason: U.S. Department of Energy guidance commonly cites leakage of 20 to 30 percent of a plant's compressed air production. Every orifice hisses in ultrasound, and the camera maps dozens of them across a compressor room and distribution network in a single walk-through. A typical survey route covers entire production halls in one shift, because each scan captures a wide field of view and the operator only closes in to tag and photograph confirmed leaks. The repairs are usually cheap, clamps and fittings, so this application frequently pays for the camera on its own.

Gas and Vacuum Leaks

The same emission principle applies to any gas escaping from a pressurized system: nitrogen, argon, natural gas lines, refrigerant circuits, and process gas. Vacuum systems work in reverse. Atmospheric air rushing into a vacuum leak generates the same turbulence and the same ultrasound, so the camera finds vacuum furnace leaks, condenser leaks, and packaging line vacuum failures without pressurizing anything. Sites with regulatory leak programs often fold these acoustic surveys into their leak detection and repair (LDAR) workflow; the Leak Detection and Repair LDAR approach covers the program and recordkeeping side of that work.

Electrical Inspections: Corona, Arcing, and Partial Discharge

Electrical faults are where the standoff distance earns its keep. Corona discharge, tracking, and arcing on switchgear, busbars, insulators, and transformers all emit concentrated ultrasound, often above 40 kHz, and the camera locates the offending component from a safe distance with no contact and no outage. Partial discharge inside insulation is only detectable acoustically once it produces a surface signature, so acoustic imaging serves as a screening and localization layer: it tells you which component to test more deeply, and its discharge signatures (corona, tracking, arcing) give a rough severity indication. It complements, rather than replaces, thermography and dielectric testing.

Mechanical Inspections: Bearings and Lubrication

As a bearing's lubrication film degrades, friction generates high-frequency noise before vibration levels rise enough to show on a vibration route. Ultrasound picks this up early, and an acoustic camera localizes which bearing in a gearbox, conveyor, or fan assembly is the noisy one, which matters on machinery where several bearings sit close together. The same logic applies to steam traps, valves, and couplings. For trendable, continuous fault detection on rotating assets, this survey capability pairs naturally with a broader Acoustic Monitoring program.

Acoustic Imaging vs Ultrasound Analysis vs Acoustic Analysis

These three terms overlap, and vendors use them loosely, so it helps to separate them cleanly.

In short: acoustic analysis is the discipline, ultrasound analysis is the measurement technique, and acoustic imaging is the technique plus a picture of the scene.

Acoustic Imaging vs Infrared vs Ultrasound Spot-Checking

Acoustic cameras share a toolkit with thermal cameras and handheld ultrasonic probes, and each finds a different slice of the fault population.

Two practical notes on the boundaries. First, infrared needs a temperature difference and a clear line of sight with manageable emissivity, so it misses cold leaks that an acoustic camera sees, and the acoustic camera misses a hot joint that thermal imaging catches immediately; mature plants run both. Second, ultrasound spot-checking relies on the operator's ear and experience for localization, so it stays strong for targeted testing on a known asset list, while the camera wins on coverage and on handover, because the finding is a photograph.

Advantages and Limitations of Acoustic Imaging

Advantages

• Standoff distance. Scans work at standoff distances of several meters up to tens of meters, which keeps people out of arc flash zones, off ladders, and away from live machinery.
• Visual localization. The overlay pins the noise source to a specific component in a photograph, which removes the guesswork of probe sweeping and produces evidence a repair technician can act on directly.
• Survey speed. Wide fields of view let one operator screen hundreds of assets in a shift, making full-plant leak routes economically realistic.
• Non-contact and non-intrusive. Nothing touches the process, no couplant, no sensors mounted, no shutdown required for a scan.
• Broad fault coverage in one tool. The same camera handles compressed air, gas, vacuum, electrical, and mechanical work by switching frequency bands.

Limitations

• Quantification is approximate. Cameras estimate leak size and cost from sound level, which is useful for prioritizing but not for billing-grade measurement. Flow meters still set the baseline.
• Noisy environments push down sensitivity. Heavy background noise masks the smallest sources, and band selection plus standoff distance become a judgment call that affects what the survey finds.
• Reflections and occlusion. Ultrasound reflects off hard surfaces, so a source behind a guard or inside a housing may appear offset from its true position or be masked by its own reflection.
• No view inside the machine. Acoustic imaging reads what escapes to the air. Internal component wear that never produces a surface signature belongs to vibration analysis and oil analysis.
• Cost and interpretation. Array cameras cost more than handheld probes, and consistent classification of electrical signatures still requires trained people.

The Bottom Line

Acoustic imaging gives ultrasound inspection the one thing a probe never had: a picture of where the sound comes from. That single change turns specialist point-checking into plant-wide surveys, and it is why the cameras have become standard equipment for compressed air, gas, vacuum, and electrical screening work. Treat it as the localization layer of a condition program. It finds and prioritizes the faults that announce themselves in air, while vibration, thermal, and oil analysis cover what it cannot see, and fixed-sensor acoustic monitoring covers the assets that cannot wait for the next survey round.

Frequently Asked Questions