Turn a phone into a scientific instrument
We develop scientific measurement applications around smartphone cameras, onboard sensors and connected hardware—combining calibration, signal processing, guided field use and production data infrastructure.
A phone is a mass-produced package of cameras, sensors, computation, connectivity and an interface people already understand. Used carefully, it can become part of a serious measurement system. Used carelessly, it produces numbers that look scientific without being reliable.
Pocket Science works in the difficult space between those outcomes.
More than reading a sensor value
A scientific measurement app may need to control acquisition conditions, interpret device metadata, compensate for hardware differences, guide physical positioning, process raw signals, detect invalid measurements and preserve enough provenance for later analysis.
- Camera-based sensing
Image acquisition, RAW data, spectral or radiometric processing and optical add-ons. - Onboard phone sensors
Location, orientation, motion, time and device-state context. - Connected instruments
Bluetooth or networked hardware, firmware-aware workflows and device diagnostics. - Calibration
Reference procedures, device profiles, quality bounds and versioned processing. - Field guidance
Instructions and feedback that help non-specialists perform a repeatable method. - Interpretable results
Outputs designed for both participants and the downstream research process.
Our development process
1. Establish the evidence boundary
We clarify what the measurement is intended to support, what is already validated and what still requires investigation. We do not manufacture equivalence claims between a phone and laboratory equipment.
2. Characterise the acquisition chain
We map the physical phenomenon, sensor, device variability, environmental conditions, calibration references, processing steps and required metadata.
3. Design the human procedure
The best algorithm cannot repair a measurement procedure that participants cannot perform. Positioning, timing, light, cleanliness, repetition and contextual observations may all belong in the interface.
4. Build processing and quality control
Depending on the method, processing can run on-device, in a backend or across both. Quality indicators and failure states are made visible instead of silently producing misleading output.
5. Validate with domain owners
Scientific authority stays with the qualified researchers. We provide the engineering, testing and traceability required to evaluate the implemented method.
Examples
iSPEX combined a smartphone, optical add-on, acquisition workflow and processing chain for spectropolarimetry. Dark Sky Meter used phone-camera data for night-sky brightness measurement. Mini Secchi combined a physical disk with guided observations of water colour and clarity.
When this service is a fit
Talk to us when your project has a real scientific method but needs a practical participant-facing instrument; when an existing sensor must become usable in the field; or when a phone camera or sensor may reduce cost and widen participation without hiding the method’s limitations.