Why it matters
Without AI, misinterpretation of magnetic anomalies can delay design decisions, miss buried features, or trigger unnecessary trenching, later risking failure to satisfy planning conditions and HER requirements.
From design to safety
Design
AI-assisted survey planning and data routing
AI ingest reads fluxgate gradiometer readings, survey grid coordinates, and historical context to propose target zones, wireframes for anomaly maps, and a provisional trench layout; it also notes data gaps, planned calibration windows, integration with the written scheme of investigation, and cross-referencing with HER records and CIfA standards.
- fluxgate gradiometers
- survey grids
- anomaly maps
- ditches
Build
Real-time data capture and QC
AI monitors live fluxgate data and grid alignment, flags drift, applies drift corrections, and updates pilot trench designs and recording plans; it records line length, offset tolerances, and situational weather impacts for later audit, and logs battery health and sensor temperature for data quality.
- fluxgate gradiometers
- survey grids
- anomaly maps
Operation
Anomaly interpretation and trench targeting
AI cross-references anomaly maps with existing context sheets, predicts likely ditch or pit features, and suggests watching briefs and immediate trenching steps; it logs feature confidence levels, sampling depth targets, tie-ins to Harris matrix, and updates interpretation plots with new layer sequences.
- anomaly maps
- ditches
- pits
Maintenance
Data curation and archival tagging
AI exports context sheets, links to interpretation plots, updates archive metadata to current schema, flags records for post-excavation assessment, generates archive deposition packets, crosswalks to HER entries, and updates file provenance records.
- interpretation plots
- anomaly maps
Safety
Hazard-aware excavation planning
AI flags potential buried services and geophysical artefacts, recommends safe spacing and monitoring during trenching per site safety plan; it integrates service maps, utility records, borehole logs, and contamination indicators for risk assessment and planning compliance.
- ditches
- pits
- interpretation plots
Part by part
- 1
fluxgate gradiometers
- Watch for
- Sensor drift or misalignment causing systematic bias in gradient readings.
- AI action
- Detect drift, trigger recalibration, re-align survey lines in real time.
- 2
survey grids
- Watch for
- Line misalignment or gaps in grid coverage compromising data continuity.
- AI action
- Flag gaps, adjust grid geometry, prompt re-survey where needed.
- 3
anomaly maps
- Watch for
- Spurious anomalies from cultural or instrumental noise.
- AI action
- Filter noise, classify genuine anomalies, surface candidates for review.
- 4
ditches
- Watch for
- Predicted trench alignments deviating from plan or unexpected feature edges.
- AI action
- Suggest trench placement adjustments, update Harris matrix links.
- 5
pits
- Watch for
- Small, isolated features near root zones or disturbances.
- AI action
- Prioritize test pits, allocate context sheets, adjust sampling plan.
- 6
interpretation plots
- Watch for
- Ambiguous feature shapes or overlapping contexts in plots.
- AI action
- Provide clarifying layers, export revised context relationships.
Worked example
Magnetic anomaly-guided trench targeting on 8.2 ha development
Illustrative scenario with sample data.
An AI-assisted workflow helps interpret magnetometry data to guide trench placement during a development-led excavation.
Signal in
input_date: 2026-03-14, survey_area_ha: 8.2, fluxgate_readings: {line1:[0.12,0.08], line2:[0.15,0.09]}, grid_coords: [(100,200),(101,200)], existing_contexts: [hedgerow trench, ancient ditch feature], anomalies_detected: [A1,A2], dataset_quality: high, calibration_window: 2026-03-13 to 2026-03-20, HER_reference: H-2026-AV18
System out
anomaly_maps: A1, A2; proposed_trenches: T5, T9; recommendation: commence watch-for and record baseline strip map.
Human sign-off: Project geophysicist reviews AI-generated targets, confirms trench locations with planning archaeologist, and approves risk assessment before field team mobilization.
How to roll it out
Phase 1
Weeks 1 to 2
Baseline data ingestion, grid setup, and initial anomaly map generation.
Phase 2
Weeks 3 to 6
Iterative interpretation refinement, trench planning, and recording plan alignment.
Phase 3
Weeks 7 to 10
Operational deployment, field targeting, and post-survey archival tagging.
Where people stay in charge
- Decisions affecting legal compliance and reporting must be signed off by the project manager or planning archaeologist with HER liaison.
- All safety-critical determinations require qualification by the site safety lead or competent person before action.
- Archive and post-excavation decisions must be validated by the finds and environmental supervisor.
Questions owners and crews ask
What data does the magnetometry AI require to start planning trenches?
The AI requires fluxgate readings, survey grid coordinates, a baseline grid map, and any existing context sheets or known features to generate anomaly maps and trench targets.
Who validates AI-produced anomaly maps before trenching?
A qualified geophysicist and the planning archaeologist validate maps, cross-checking with written scheme of investigation and HER data before fieldwork proceeds.
How does AI handle noise in magnetic data during build and operation?
The system applies filters, classifies signals, and flags potential noise sources, prompting review and, if needed, re-collection of affected lines.
What records are created during interpretation plotting?
Interpretation plots link context sheets, Harris matrix entries, anomaly surfaces, and trench plans, with metadata and audit trails for post-excavation assessment.