Mine blast event detection and production rhythm mapping
Correlating high-revisit optical imagery with public seismic and infrasound catalogues lets analysts time blast events, map bench exposure, and infer production rhythms at open-pit mines anywhere on Earth.
Sensors
- Planet SuperDove: 3 m ground sample distance, 8 spectral bands including a red-edge channel, daily global revisit under the full constellation. The primary optical clock for detecting plume onset and fresh bench exposure within a 24-hour window of a blast event.
- Planet SkySat: 50 cm resolution, taskable to a specific pit on demand. Used to confirm bench geometry and exposed-rock extent after a blast flagged by SuperDove, and to resolve individual blast-hole patterns on large benches.
- Sentinel-2 MSI: 10 m resolution in visible and near-infrared bands, 5-day revisit at the equator (better at higher latitudes). Free archive from 2015. Useful for multi-year production rhythm baselines and for cross-validating SuperDove detections.
- Sentinel-1 SAR (coherence loss): C-band synthetic aperture radar, 10 m resolution in Interferometric Wide Swath mode, 6-day repeat. Ground disturbance from blasting destroys interferometric coherence between repeat passes. A coherence-drop map localises the blast zone independently of cloud cover or dust opacity.
What a blast actually looks like from 500 km up
A production blast at an open-pit mine is a discrete, repeatable event with a predictable spectral and textural signature. In the seconds after detonation, a dust plume rises from the bench. In the minutes that follow, freshly fractured rock is exposed: spectrally distinct from weathered overburden because its surface has never seen sunlight, rain or oxidation. Both signatures are detectable in multispectral imagery if the revisit is tight enough.
Planet SuperDove's daily cadence means that, under clear skies, a blast occurring on day N is visible in imagery acquired on day N or N+1. The 8-band sensor captures the near-infrared reflectance jump that accompanies fresh rock exposure, and the visible-band brightness anomaly of the plume itself if the acquisition passes within hours of the event. At 3 m resolution, bench-scale geometry is legible: a blasted bench typically spans tens to hundreds of metres along strike, well above the detection floor.
The seismic and infrasound record as a ground truth
Mine blasts are among the most reliably catalogued seismic sources on Earth. National and regional seismic networks routinely discriminate quarry and mine blasts from tectonic events using waveform characteristics: the surface-wave to body-wave ratio, the P-wave first-motion pattern, and the time-of-day clustering that distinguishes industrial blasting from natural seismicity. The International Seismological Centre publishes a Bulletin of detected events, and the USGS National Earthquake Information Center maintains a publicly searchable catalogue. Many large mines file blast notifications with national regulators, creating a paper record that can be cross-referenced.
Infrasound, the sub-20 Hz acoustic signal produced by large surface blasts, propagates hundreds of kilometres and is monitored by the International Monitoring System operated under the Comprehensive Nuclear-Test-Ban Treaty Organisation. IMS stations are not designed for mine monitoring, but their open data have been used in published research to detect and locate large industrial blasts at ranges exceeding 1,000 km. The combination of a seismic or infrasound timestamp with a same-day satellite image is what transforms a spectral anomaly into a confirmed blast event.
Building a production rhythm from individual events
A single confirmed blast is an observation. A time series of confirmed blasts is a production clock. Open-pit mines blast on schedules driven by shift patterns, regulatory restrictions on night blasting, and the operational logic of the drill-blast-load-haul cycle. Most large operations blast once or twice per working day, often in the late morning after the drilling shift. That regularity means a 90-day archive of daily imagery, cross-referenced against the public seismic catalogue, can reconstruct the blast schedule with enough fidelity to infer working days, maintenance shutdowns, and periods of curtailed production.
Sentinel-1 coherence loss adds a cloud-independent layer. When a blast disturbs the ground surface between two SAR acquisitions six days apart, the interferometric coherence in that pixel drops sharply, typically below 0.3 in C-band, against a stable background of 0.6 to 0.9 over undisturbed rock. The spatial pattern of coherence loss maps the blast footprint even when optical imagery is obscured. Combining optical event timestamps with SAR footprint maps produces a spatial-temporal record: which benches were blasted, in what sequence, and at what cadence.
There are honest limits. Cloud cover breaks the optical record; equatorial and tropical mines may have weeks of unusable imagery during wet seasons. The seismic catalogues have detection thresholds: small blasts below roughly magnitude 1.0 to 1.5 local magnitude may not appear. And the method cannot distinguish a blast from a large rockfall without corroborating evidence. Where the optical and seismic signals align, the inference is strong. Where only one signal is present, confidence is lower and the deliverable should say so.
Coherence loss as a bench-mapping tool
The Sentinel-1 coherence approach deserves its own explanation because it is often underused. When rock is blasted, the surface is physically rearranged at centimetre scale. C-band radar, with a wavelength of roughly 5.6 cm, is sensitive to surface changes at that scale. A pixel that was coherent across two passes, meaning the radar scatterers had not moved, becomes incoherent after blasting. The loss is not subtle: it is typically the dominant coherence signal in the pit area during an active production period.
By differencing consecutive coherence maps, an analyst can identify which spatial units within the pit were disturbed in each 6-day interval. Stacked over a year, this produces a bench-advance sequence: a map of where the mine has been working, in what order, and at roughly what pace. This is independent of spectral analysis entirely, which makes it a useful cross-check and a fallback when optical data are unavailable.
What this analysis can and cannot tell a buyer
Production rhythm mapping from blast detection is a proxy, not a direct measurement. It tells you when and where blasting occurred, and at what cadence. It does not tell you the tonnage blasted, the ore grade, or the recovery rate at the processing plant. Inferring tonnes requires assumptions about bench geometry and rock density that introduce substantial uncertainty unless validated against disclosed production data.
The method is most valuable for competitive intelligence on undisclosed operations, for supply-chain due diligence where a counterparty's production claims need independent verification, and for regulatory monitoring of mines operating under blast-frequency restrictions. It is also useful for tracking recovery after a shutdown: the return of regular blast signatures in the seismic catalogue, confirmed by fresh bench exposure in optical imagery, is a reliable signal that a mine has resumed production.
Satellize runs this analysis on open Sentinel-1 and Sentinel-2 archives combined with commercial SuperDove tasking, applying the same change-detection pipeline used in its Tonga crop-estimation programme to a very different physical signal. The method is transferable; the physics is the same.
Archive depth and the value of the long record
Sentinel-1 data run back to 2014 for many sites. Sentinel-2 covers from 2015. Planet's SuperDove archive is shallower, from around 2021 for full 8-band coverage, but the earlier 4-band Dove archive extends further. This means a decade-scale production rhythm analysis is possible for many major mines, using Sentinel as the backbone and commercial data for higher-resolution event confirmation.
A long archive matters because it captures the full operating cycle: ramp-up, steady state, curtailment, and closure. A mine that claims continuous operation but shows a two-year gap in blast signatures has a story to tell. Conversely, a mine that is accelerating its blast cadence beyond its permitted schedule is visible in the same record. The archive is the audit trail.
Typical figures
| Primary optical resolution | 3 m (Planet SuperDove); 50 cm on-demand (Planet SkySat); 10 m (Sentinel-2 MSI) |
| Optical revisit | Daily under full Planet constellation; 5-day at equator for Sentinel-2 (shorter at higher latitudes) |
| SAR coherence resolution | 10 m (Sentinel-1 IW mode); 6-day repeat baseline |
| Seismic catalogue latency | USGS and ISC preliminary bulletins typically within hours to days of event |
| Minimum detectable blast (seismic) | Approximately Ml 1.0 to 1.5 for regional networks; larger for IMS infrasound (site-dependent) |
| Spectral bands used | Visible (blue, green, red), red-edge, near-infrared (SuperDove 8-band); C-band 5.6 cm (Sentinel-1) |
| Archive depth | Sentinel-1 from 2014; Sentinel-2 from 2015; SuperDove 8-band from approx. 2021 |
| Cloud sensitivity | Optical methods fail under cloud; SAR coherence is cloud-independent and provides fallback |
| Typical analysis latency | 24 to 72 hours from image acquisition to confirmed event report, depending on seismic catalogue update cycle |
| Delivery formats | GeoTIFF coherence-loss layers, GeoJSON event point catalogue, PDF rhythm report, time-series CSV |
Analytics Satellize can run
| Blast event catalogue | Optical plume and fresh-exposure detection (spectral change thresholding on SuperDove NIR and red-edge bands) cross-referenced against USGS or ISC seismic event timestamps | GeoJSON point catalogue with event date, location, confidence score, and source signal (optical, seismic, or both) |
| Bench disturbance footprint map | Sentinel-1 interferometric coherence differencing between consecutive 6-day repeat passes; coherence-drop threshold applied at pixel level | GeoTIFF coherence-loss layer per acquisition interval, stacked into a quarterly disturbance mosaic |
| Production rhythm summary | Time-series analysis of blast event frequency and inter-event intervals; working-day and shutdown-period classification | Monthly PDF report with blast-frequency chart, identified shutdown periods, and cadence trend over the analysis window |
| Bench advance sequence map | Spatial clustering of coherence-loss footprints over a rolling 12-month window to reconstruct the order and pace of bench development | Annotated GeoTIFF showing bench-advance chronology; compatible with standard GIS platforms |
| Production resumption alert | Automated detection of first blast event following a silence period exceeding a user-defined threshold (e.g. 14 days); triggered by seismic catalogue update and confirmed by next available optical pass | Email or API alert with supporting image chip and seismic event reference |
| Multi-year production rhythm baseline | Sentinel-1 and Sentinel-2 archive analysis from 2014/2015 to present; blast cadence reconstructed at monthly resolution using coherence loss and spectral change time series | Time-series dataset (CSV) and summary report covering full archive period, with annotated periods of curtailment or acceleration |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.