Landslide Detection Drone Guide

Landslides are a significant natural hazard affecting communities, transport networks, construction projects, mines, utilities and critical infrastructure around the world. They can develop gradually over months or occur suddenly following heavy rainfall, snowmelt, erosion, earthquakes, excavation or changes in groundwater conditions. Monitoring potentially unstable slopes has traditionally relied on ground surveys, fixed instrumentation, satellite observations and visual inspections, but drones have become an increasingly valuable addition to this monitoring network.

Drones can rapidly collect high-resolution information across slopes that may be difficult or dangerous for survey teams to access. Using combinations of RGB cameras, LiDAR, thermal cameras, multispectral sensors, photogrammetry and precise positioning systems, drones can document terrain, identify visible changes and support repeated measurements of potentially unstable areas.

Their greatest value often comes from repeatability. A single flight provides a detailed snapshot of a slope, but repeated surveys allow specialists to compare terrain over time. Small changes in surface geometry, cracks, displaced material, drainage patterns or vegetation may indicate that further geotechnical investigation is required.

However, drones do not independently determine whether a landslide will occur. Surface movement can be an important indicator of instability, but landslides are complex geotechnical processes involving geology, groundwater, soil properties, slope geometry and loading conditions. Drone observations should therefore complement professional geotechnical assessment and ground-based monitoring rather than replace them.

Understanding Landslide Detection with Drones

The term landslide detection can describe several different activities. A drone may be used to identify an existing landslide, map the extent of a recent event, monitor a known unstable slope or identify surface changes that may indicate developing instability.

These activities should be distinguished from predicting exactly when a landslide will occur. A drone can measure visible or geometric changes, but those observations alone usually cannot determine the timing of future slope failure.

The strongest monitoring programmes combine drone measurements with information from instruments such as inclinometers, piezometers, extensometers, GNSS monitoring stations, weather stations and groundwater sensors. Geological mapping and engineering analysis provide additional context.

Drones therefore become another layer within a broader landslide monitoring system.

Why Drones Are Valuable for Landslide Monitoring

Potentially unstable slopes can be difficult to inspect from the ground. Steep terrain, loose material, rockfall hazards, vegetation and poor access can expose survey teams to unnecessary risk.

A drone allows much of the initial observation to be performed remotely.

Large slopes can also be surveyed much more comprehensively than with individual ground measurements. Instead of collecting information from a limited number of locations, photogrammetry or LiDAR can produce millions of measurements covering much of the visible surface.

This creates a detailed baseline against which future surveys can be compared.

Drones are particularly valuable following heavy rainfall or an initial slope movement when authorities need current information but entering the affected area may be unsafe.

RGB Cameras

High-resolution RGB cameras are among the simplest and most useful sensors for landslide monitoring.

Detailed photographs can reveal cracks, exposed soil, displaced rocks, damaged vegetation, erosion channels and changes in drainage. Orthomosaics can provide a georeferenced overview of the complete area.

Oblique imagery is particularly useful on steep slopes because a conventional downward-facing camera may not adequately capture near-vertical terrain.

RGB imagery also provides important context for measurements from other sensors.

However, visible appearance alone cannot determine whether a slope is stable. A crack may indicate movement, but its significance depends on its location, geometry, development and underlying geology.

Drone Photogrammetry

Photogrammetry converts overlapping drone photographs into three-dimensional information.

Structure-from-Motion processing can generate dense point clouds, Digital Surface Models, orthomosaics and textured 3D models.

For landslide monitoring, this provides a relatively accessible method of measuring slope geometry.

Repeated photogrammetric surveys can be compared to identify changes in elevation and surface position.

Areas where material has been removed can be distinguished from areas where material has accumulated.

However, reliable change detection requires accurate alignment between surveys. Differences caused by poor georeferencing should not be mistaken for actual ground movement.

LiDAR for Landslide Detection

LiDAR is particularly valuable for landslide monitoring because it directly measures three-dimensional geometry.

A drone-mounted LiDAR scanner can collect dense point clouds of slopes, cliffs and surrounding terrain. Multiple laser returns may also provide ground measurements through gaps in vegetation.

This can make LiDAR especially useful in forested or partially vegetated landslide areas.

Repeated LiDAR surveys can reveal changes in slope geometry, erosion and displaced material.

However, LiDAR should not be described as simply seeing through vegetation. Laser pulses reach the ground through openings in the canopy. Extremely dense vegetation may still prevent adequate terrain measurement.

Digital Terrain Models

One of the most useful products for landslide analysis is a Digital Terrain Model.

A DTM attempts to represent the ground after vegetation, buildings and other objects have been removed from the point cloud.

This can reveal terrain features associated with previous or ongoing slope movement.

Scarps, depressions, displaced blocks, drainage channels and accumulation zones may become easier to recognise.

Hillshade, slope and curvature models derived from the DTM can further support interpretation.

However, terrain morphology alone does not prove that a landslide is currently active.

Establishing a Baseline Survey

Effective change detection begins with a reliable baseline.

The first drone survey should capture the complete monitoring area with sufficient detail and accuracy.

This baseline becomes the reference against which later flights are compared.

Survey control should be established where the required accuracy justifies it. RTK or PPK positioning can improve georeferencing, while independently surveyed check points provide evidence of actual accuracy.

The baseline should also include surrounding stable terrain where possible. These areas can help verify that apparent changes between surveys are genuine rather than alignment errors.

Repeat Surveys and Change Detection

The real strength of drone landslide monitoring appears when surveys are repeated.

A slope may be flown weekly, monthly, seasonally or after specific triggering events such as heavy rainfall.

Point clouds or terrain models from different dates can then be aligned and compared.

Software can calculate surface-to-surface differences, highlighting where terrain has moved, eroded or accumulated.

For example, an upper slope may show loss of material while a lower section shows deposition.

The magnitude and pattern of these changes can help geotechnical specialists understand how the slope is evolving.

Detecting Surface Movement

Surface movement may appear as horizontal displacement, vertical change or a combination of both.

Drone photogrammetry and LiDAR can support measurement of these changes where identifiable features remain visible between surveys.

However, not every difference represents ground movement.

Vegetation growth, vehicles, construction activity, snow and processing differences can all create apparent changes.

Professional workflows therefore separate genuine terrain change from temporary or irrelevant differences before interpreting the results.

Detecting Cracks and Fissures

Cracks can be important visible indicators of slope deformation.

High-resolution RGB imagery can document their location and development.

A crack may be mapped in an orthomosaic and revisited during subsequent surveys.

Very large cracks may also appear in LiDAR geometry.

However, detecting a crack from imagery does not establish its depth or geotechnical significance.

Some cracks may result from drying, erosion or local surface processes rather than deep-seated slope movement.

Ground inspection may therefore be required.

Monitoring Head Scarps

A head scarp is the upper boundary created when part of a slope moves downward.

Existing scarps can be mapped accurately with LiDAR or photogrammetry.

Repeat surveys may show whether the scarp is retreating or expanding.

Changes around the upper part of a landslide can be particularly important for understanding whether additional material may become involved.

However, drone measurements describe visible geometric change. Geotechnical specialists must interpret what that change means for future slope behaviour.

Monitoring the Landslide Toe

The toe is the lower portion of displaced landslide material.

Drone surveys can map its extent and identify changes over time.

This can be particularly important where the toe approaches roads, buildings, rivers or other infrastructure.

Repeated terrain models can show whether material is advancing.

However, surface geometry does not reveal all movement occurring within the landslide mass.

Subsurface instrumentation may still be necessary.

Rockfall and Cliff Monitoring

LiDAR drones can map cliffs and rock faces in three dimensions.

Repeat point clouds can identify where rock has detached.

High-resolution RGB imagery may reveal fractures and weathering.

This can support rockfall hazard assessment along roads, railways, quarries and mountain infrastructure.

Oblique flight paths are often more useful than purely vertical mapping because they capture the face directly.

However, visible fractures do not by themselves determine when a rock block will fail.

Engineering geology remains essential.

Rainfall-Triggered Landslides

Heavy rainfall is one of the most common landslide triggers.

Water can infiltrate soil and increase pore-water pressure, reducing effective strength within the slope.

After major rainfall, drones can rapidly inspect known risk areas for new cracks, erosion or movement.

Automated monitoring programmes could also trigger drone missions when rainfall thresholds are exceeded.

However, drone imagery cannot directly measure pore-water pressure.

Piezometers and hydrological monitoring remain important for understanding subsurface conditions.

Groundwater and Drainage

Changes in groundwater can have a major influence on slope stability.

Drone imagery can help identify surface water, blocked drainage, new channels or areas of persistent wetness.

Thermal and multispectral observations may sometimes provide additional indications of moisture differences.

These observations can guide field investigation.

However, surface moisture should not be treated as a direct measurement of groundwater conditions.

A dry-looking surface does not mean that the underlying slope is dry.

Thermal Imaging

Thermal cameras measure differences in surface temperature.

In some environments, these differences may help identify drainage patterns, seepage or areas with different moisture characteristics.

Thermal surveys can therefore complement RGB and LiDAR information.

However, thermal patterns are influenced by sunlight, shade, vegetation, material properties, wind and time of day.

A thermal anomaly does not automatically indicate groundwater or instability.

The strongest thermal surveys are planned for environmental conditions that maximise useful temperature contrast and are interpreted alongside other information.

Multispectral Imaging

Multispectral cameras measure reflected energy across selected wavelength bands.

Vegetation indices can be generated to examine changes in plant condition.

Vegetation stress may sometimes correspond with changing drainage, soil movement or disturbed ground.

Multispectral imagery can also help classify land cover.

However, vegetation stress has many possible causes, including drought, disease and nutrient deficiency.

It should therefore be treated as supporting evidence rather than proof of slope instability.

Vegetation as an Indicator

Vegetation can provide useful visual clues about slope movement.

Tilted trees, disturbed vegetation or newly exposed soil may indicate recent change.

Historical aerial imagery can sometimes reveal long-term patterns.

However, vegetation can also obscure the ground.

Combining RGB imagery with LiDAR is therefore valuable because visual observations can be compared with underlying terrain geometry.

Professional interpretation should consider both biological and geological explanations for observed changes.

Landslide Volume Calculation

Drone-derived 3D models can be used to estimate the volume of displaced material.

Pre-event and post-event terrain models provide the strongest basis because they show the surface before and after failure.

Where no pre-event drone survey exists, historical LiDAR or other terrain data may provide a reference.

Volume calculations can help authorities understand the scale of an event and estimate material-removal requirements.

However, the calculation represents differences between available surfaces and may not capture all subsurface deformation.

Roads and Highways

Road corridors are highly exposed to landslide risk in mountainous terrain.

Drones can inspect slopes above and below roads without requiring survey teams to climb unstable areas.

LiDAR and photogrammetry can map cut slopes, retaining structures, drainage and surrounding terrain.

Repeat surveys can identify significant geometric changes.

Following an event, drones can also map debris covering the road and provide information for clearance planning.

However, a visually clear road should not be assumed safe to reopen until appropriate engineering and operational assessments are complete.

Railway Landslide Monitoring

Railways can be vulnerable to rockfalls, embankment failures and slope movement.

Drone surveys can monitor cuttings and slopes along the corridor.

LiDAR is particularly useful for creating detailed terrain models.

Repeat surveys can highlight changes near tracks.

However, railway safety decisions require specialist engineering and operational procedures.

Drone information supports those decisions but should not independently determine whether a railway is safe for traffic.

Pipelines

Pipelines crossing mountainous or unstable terrain may be affected by landslides.

Drone LiDAR can map the terrain surrounding the pipeline corridor.

Repeat surveys can identify major slope deformation, erosion or ground movement.

RGB imagery provides additional visual evidence.

However, the movement of buried pipelines cannot normally be determined directly from surface LiDAR alone.

Pipeline instrumentation and geotechnical monitoring may therefore need to be integrated.

Powerlines

Landslides can threaten towers, poles and access roads.

Drone surveys can map both the terrain and the utility infrastructure.

LiDAR provides the three-dimensional relationship between slopes and assets.

Repeat surveys can identify terrain changes near foundations.

However, the presence of a landslide near a tower does not automatically establish that the foundation has been structurally compromised.

Engineering inspection remains necessary.

Construction Sites

Excavation and earthworks can alter slope geometry and drainage.

Drones can provide frequent topographic surveys during construction.

This allows engineers to compare actual terrain with design surfaces and identify unexpected changes.

Large cut slopes can be monitored without requiring surveyors to access every section.

However, drone measurements do not replace geotechnical instrumentation where slope stability is a significant project risk.

Open-Pit Mines

Open-pit mines contain large engineered slopes that change continuously.

Drone LiDAR and photogrammetry can provide frequent high-resolution models of benches and pit walls.

Surface changes can be identified between surveys.

Drones can reduce the need for personnel to approach potentially unstable areas.

However, mine slope management normally combines remote sensing with radar, prisms, geotechnical instrumentation and geological observations.

Drone surveys provide an additional spatial layer within that monitoring system.

Quarries

Quarry faces can be mapped with oblique photogrammetry or LiDAR.

Repeat surveys can document excavation and identify areas of surface change.

This information can support geological and operational assessment.

However, quarry activity itself creates major terrain differences.

Processing workflows need to distinguish intentional excavation from unexpected slope movement.

Operational records therefore provide valuable context.

Dams and Reservoirs

Slopes surrounding reservoirs may be affected by changing water levels, erosion and geological conditions.

Drones can monitor exposed terrain and identify visible changes.

LiDAR can create accurate terrain models, while RGB imagery documents surface condition.

However, reservoir-related slope stability may involve deep groundwater and geological processes that cannot be observed directly from the air.

Drone information should therefore be integrated with geotechnical and hydrological monitoring.

Coastal Landslides

Coastal cliffs are continuously affected by waves, rainfall and weathering.

Drones provide an effective method for repeatedly mapping these difficult-to-access areas.

LiDAR and photogrammetry can quantify cliff retreat.

After a collapse, new surveys can calculate the volume of material lost.

Bathymetric LiDAR or other marine surveying methods may extend the analysis into shallow water.

However, cliff failure can occur suddenly, so appropriate stand-off distances should be maintained during drone operations.

Riverbank Instability

River erosion can undermine banks and trigger slope failures.

Drone surveys can map bank geometry and identify areas of retreat.

Repeat models can show how the channel and surrounding slopes are changing.

Where water clarity permits, bathymetric LiDAR may provide additional information about the submerged channel.

However, conventional aerial LiDAR does not reliably measure underwater terrain.

Hydraulic and geotechnical information may therefore need to be combined.

Forested Slopes

Forested terrain is particularly challenging for conventional image-based mapping because vegetation obscures the ground.

LiDAR offers an important advantage because some laser pulses can pass through gaps in the canopy.

Ground-classified point clouds can reveal slope morphology beneath vegetation.

This may expose historical landslide features that are difficult to recognise from ordinary photographs.

However, dense canopy can still limit ground returns.

The resulting terrain model should therefore be checked for interpolation and coverage gaps.

Snow and Mountain Environments

Mountain landslides can interact with snow, freeze-thaw processes and seasonal runoff.

Drones can provide rapid terrain surveys during accessible weather windows.

However, snow changes the observed surface.

Comparing a snow-covered survey with a snow-free terrain model may create large apparent differences unrelated to ground movement.

Repeat monitoring should therefore account for seasonal surface conditions.

Earthquake-Triggered Landslides

Earthquakes can trigger large numbers of landslides across extensive regions.

Drones can provide detailed local surveys of high-priority sites after the event.

They may help map blocked roads, damaged slopes and debris.

Satellite imagery can first identify broad affected areas, with drones then collecting higher-resolution information.

This satellite-to-drone workflow can be particularly efficient.

However, aftershocks and unstable terrain may continue to create hazards, so drone deployment should follow emergency-management priorities.

Post-Landslide Emergency Assessment

Following a landslide, emergency teams need to understand the extent of the event quickly.

Drones can map the source area, debris path and accumulation zone without immediately sending personnel onto unstable ground.

RGB cameras provide visual situational awareness, while LiDAR or photogrammetry provides measurable geometry.

Thermal cameras may support search activities under suitable circumstances.

However, the absence of a thermal detection does not establish that no person is present.

Search-and-rescue professionals should interpret all sensor information within the broader response.

Search and Rescue

Drones may support search operations following major landslides.

RGB and thermal cameras can scan debris and surrounding terrain.

LiDAR can provide a 3D model that helps teams understand access routes and debris geometry.

However, thermal detection can be limited by burial depth, environmental temperature, debris and vegetation.

A thermal anomaly may indicate many different heat sources.

Drone observations should therefore support rather than replace established search-and-rescue methods.

Mapping Landslide Boundaries

Accurate mapping of the affected area is important for emergency management, engineering and insurance.

Drone orthomosaics allow the landslide boundary to be digitised at high resolution.

LiDAR provides three-dimensional information about scarps and debris.

GIS can then combine these observations with roads, buildings, utilities and property information.

The mapped boundary should distinguish between directly observed terrain change and areas inferred from interpretation.

GIS Integration

GIS is central to professional landslide monitoring.

Drone point clouds, orthomosaics, terrain models and detected changes can be combined with geology, rainfall, infrastructure and historical landslide inventories.

This provides a broader spatial picture.

Risk managers can identify which assets lie near changing slopes.

However, proximity does not automatically equal risk.

Hazard, exposure and vulnerability need to be evaluated together by appropriate specialists.

Historical Data

Historical terrain information can provide valuable context.

Old aerial photographs, satellite imagery and previous LiDAR surveys may show how a slope has evolved over years or decades.

Current drone surveys can be compared with these datasets.

However, differences in accuracy and resolution should be considered.

A modern centimetre-resolution drone survey should not be compared directly with an older low-resolution model without understanding the uncertainty in both datasets.

Satellite Integration

Satellite monitoring can cover much larger areas than drones.

Optical satellites may identify major terrain changes, while satellite radar interferometry can measure certain patterns of ground deformation over wide areas.

Drones then provide much higher-resolution local information.

This creates a useful monitoring hierarchy:

satellite regional observation → identification of priority areas → targeted drone survey → detailed geotechnical investigation → ongoing monitoring.

The technologies complement rather than replace one another.

InSAR and Drone Surveys

Interferometric Synthetic Aperture Radar can measure ground displacement using repeated radar observations from satellites or other platforms.

It can be extremely valuable for identifying broad patterns of movement.

Drone LiDAR or photogrammetry can then provide detailed three-dimensional mapping of selected locations.

However, the two technologies measure movement differently and have different limitations.

Combining them can provide a stronger understanding than relying on either dataset alone.

Ground-Based Monitoring

Drone surveys are most powerful when integrated with ground instrumentation.

Inclinometers can measure subsurface deformation.

Piezometers monitor groundwater pressure.

GNSS stations measure precise movement at specific locations.

Extensometers measure changes across cracks.

Weather stations record rainfall and environmental conditions.

The drone provides wide-area spatial context around these point measurements.

Together, they can create a much more complete monitoring system.

Fixed GNSS Monitoring

Permanent GNSS receivers can continuously measure movement at selected locations.

They provide excellent temporal resolution but only at the locations where sensors are installed.

Drone surveys provide much broader spatial coverage but are normally periodic.

Combining the two allows continuous monitoring points to be interpreted within a detailed 3D terrain model.

This is particularly valuable for large unstable slopes.

Ground-Based Radar

Slope-monitoring radar can detect small surface movements continuously across large areas.

It is widely used in mining and other high-risk slope environments.

Drone LiDAR provides complementary high-resolution geometry and imagery.

Radar may indicate that an area is moving, while the drone helps document its physical characteristics.

Neither technology needs to replace the other.

Integrated monitoring can provide redundancy and additional context.

Ground-Penetrating Radar

Ground-penetrating radar can investigate certain subsurface conditions, depending on material properties.

It addresses a different problem from aerial LiDAR.

LiDAR measures visible surface geometry.

GPR investigates beneath the surface.

Where geotechnical conditions allow, the two technologies can complement one another.

However, aerial drones should not be assumed capable of determining deep slip surfaces simply because they carry remote-sensing equipment.

Rainfall Sensors and Weather Data

Rainfall thresholds are often used as part of landslide warning systems.

Weather stations and rainfall radar can identify when conditions associated with increased landslide likelihood are developing.

This information can trigger additional drone surveys.

A monitoring programme might conduct routine monthly flights but add flights after unusually heavy rainfall.

The resulting terrain data can then be compared with the baseline.

This creates a more responsive monitoring strategy.

Automated Drone Monitoring

Drone-in-a-Box technology could make repeated landslide monitoring more practical.

A permanently installed drone could automatically inspect a known slope after predefined events or on a regular schedule.

Consistent flight paths would improve comparison between surveys.

Data could be uploaded automatically for processing.

However, automated operation still requires reliable positioning, weather limits, airspace compliance and quality assurance.

The system should also recognise when conditions are unsuitable for useful data collection.

AI-Assisted Landslide Detection

Artificial intelligence can help analyse large volumes of drone data.

Computer-vision models may identify cracks, exposed soil or vegetation change.

Point-cloud algorithms can detect geometric differences between surveys.

AI can also prioritise areas where change exceeds predefined thresholds.

However, these outputs should be treated as candidate observations.

AI does not independently determine slope stability or predict failure.

Geotechnical professionals should interpret the results alongside geological and monitoring information.

Automated Change Detection

One of the strongest uses of automation is comparing repeated 3D datasets.

Software can calculate the distance between point clouds or terrain models and highlight areas where change exceeds a threshold.

This can rapidly narrow a very large survey down to a few areas requiring review.

However, the threshold must account for measurement uncertainty.

If the survey accuracy is several centimetres, a detected change of similar magnitude may not be meaningful.

Independent accuracy verification is therefore essential.

Machine Learning and Landslide Inventories

AI can assist with identifying terrain patterns associated with historical landslides across large datasets.

Digital terrain models can be analysed for scarps, hummocky terrain and other morphological features.

This can support landslide inventories.

However, terrain morphology can have multiple geological explanations.

Automated classification should therefore be checked by experienced geomorphologists or geotechnical specialists.

The goal is faster screening rather than replacing professional mapping.

RTK and PPK Positioning

Accurate repeat surveys require strong georeferencing.

RTK and PPK GNSS can provide centimetre-level positioning under suitable conditions.

This reduces differences caused by inaccurate aircraft location.

However, RTK or PPK alone does not guarantee that two point clouds are perfectly aligned.

Camera calibration, LiDAR navigation quality, control and processing also matter.

Independent check points remain valuable for high-accuracy change detection.

Ground Control Points

Ground Control Points can improve the absolute positioning of photogrammetric models.

They also provide stable references between repeated surveys.

Control should preferably be located on stable terrain outside the active landslide.

A control point that moves with the slope cannot provide a reliable reference.

Permanent survey monuments may therefore be valuable for long-term monitoring.

Their stability should itself be verified periodically.

Check Points

Independent check points allow survey accuracy to be tested.

They should not be used to adjust the model if they are intended as independent verification.

Differences between measured and modelled coordinates provide evidence of actual survey performance.

This is especially important when attempting to detect small changes.

A landslide monitoring programme should know the smallest movement it can reliably distinguish from survey noise.

Minimum Detectable Change

Not every measured difference between surveys represents genuine movement.

Every survey contains uncertainty.

The minimum detectable change should therefore be related to the combined uncertainty of the compared datasets.

Large terrain movements may be obvious.

Small centimetre-scale differences require much stronger control and processing.

Monitoring programmes should establish realistic detection thresholds rather than treating every coloured pixel on a change map as meaningful ground deformation.

3D Point-Cloud Comparison

Point clouds from different dates can be compared directly.

Algorithms calculate the distance between surfaces.

This can reveal where a slope has moved.

Three-dimensional comparison is particularly valuable on steep slopes where vertical-only terrain-model differencing may not represent the direction of movement well.

However, vegetation and temporary objects can create false differences.

Appropriate classification and filtering are therefore necessary.

Digital Elevation Model Differencing

Digital Elevation Models from different dates can be subtracted.

Positive values indicate apparent elevation gain, while negative values indicate apparent loss.

This provides an intuitive way to visualise erosion and deposition.

It is especially useful for calculating landslide volume.

However, DEM differencing mainly represents vertical change.

Complex horizontal movement may require point-cloud or feature-based analysis.

Feature Tracking

Visible objects or surface patterns can sometimes be tracked between images.

Their displacement provides information about movement direction.

Computer vision may automate this process.

However, features can change because of vegetation, lighting or erosion.

Tracking should therefore be validated carefully.

Stable reference areas help distinguish genuine ground movement from image-registration errors.

3D Displacement Vectors

Advanced monitoring may estimate movement in three dimensions rather than only elevation change.

This can provide information about both magnitude and direction.

LiDAR point clouds, photogrammetric models and fixed targets can contribute.

However, high-quality 3D displacement measurement requires strong geometric control.

The uncertainty should always be reported alongside the calculated movement.

Digital Twins for Landslide Monitoring

A digital twin can combine the current 3D terrain with monitoring information.

Drone LiDAR provides geometry.

Weather stations provide rainfall.

Piezometers provide groundwater information.

GNSS sensors provide movement.

Historical surveys show previous terrain.

The resulting environment gives engineers a central platform for understanding slope behaviour.

However, a digital twin is only useful if its data is current and correctly referenced.

Risk Mapping

Drone-derived terrain can contribute to landslide hazard and risk maps.

Slope angle, drainage, geology, vegetation and historical movement can be analysed together.

Infrastructure and buildings can then be overlaid.

However, a drone survey alone does not establish landslide risk.

Risk assessment requires consideration of the probability and potential consequences of failure.

Professional geological and engineering analysis remains necessary.

Emergency Route Planning

After a landslide, drone maps can help emergency teams understand blocked roads and terrain.

Aerial imagery can identify possible access routes.

LiDAR can provide slope and elevation information.

However, a route that appears physically open from the air should not automatically be considered safe.

Additional landslides, unstable debris or damaged structures may create hazards that cannot be determined from imagery alone.

Ground verification remains important.

Safety Benefits

One of the strongest reasons for using drones is reducing human exposure.

Surveyors no longer need to climb every unstable slope to collect measurements.

Emergency responders can obtain an initial overview before entering affected areas.

Mining teams can inspect high walls remotely.

Transport operators can examine cuttings without immediately placing personnel beneath potentially unstable rock.

Drones therefore provide both measurement capability and an important stand-off safety advantage.

Operational Challenges

Mountainous terrain creates demanding flight conditions.

Wind can vary significantly around ridges and valleys.

GNSS signals may be blocked by cliffs.

Communications can be lost behind terrain.

Battery performance may decrease in cold environments.

Take-off locations can also be difficult to find.

Mission planning should therefore consider both survey quality and aviation safety.

GNSS-Degraded Environments

Steep cliffs and valleys can reduce satellite visibility.

This may affect RTK or PPK positioning.

LiDAR-inertial or visual-inertial navigation can provide additional support on some platforms.

However, GNSS-denied navigation should not automatically be assumed to provide survey-grade global coordinates.

External control may still be required.

The navigation strategy should match the accuracy needed for change detection.

Weather

Landslide monitoring often becomes most important during poor weather.

Unfortunately, heavy rain, strong wind, fog and low cloud can prevent safe drone operations.

LiDAR can operate without sunlight but may still be affected by rain or fog.

RGB photogrammetry requires adequate visibility.

Monitoring programmes should therefore include alternative data sources for periods when drones cannot fly.

Fixed sensors provide important continuity.

Data Management

Repeated landslide surveys can create very large datasets.

Point clouds, imagery and terrain models should be organised consistently.

Each survey should include date, coordinate system, processing settings, control information and environmental conditions.

This metadata is important because a monitoring programme may continue for years.

Without consistent records, comparing historical surveys becomes more difficult.

Data Quality and Professional Interpretation

Drone data can be visually impressive, but visual detail should not be confused with geotechnical certainty.

A detected terrain change indicates that the observed surface appears to have changed.

It does not automatically explain why.

A crack indicates surface separation but does not automatically reveal the depth of the failure mechanism.

A wet thermal anomaly may indicate moisture but does not independently prove groundwater-driven instability.

A lack of visible movement does not prove that the slope is stable.

Professional interpretation is therefore fundamental.

Selecting a Drone System for Landslide Detection

The best drone depends on the environment and monitoring objective.

Multirotors are well suited to steep slopes and smaller sites because they can hover and collect oblique imagery.

Fixed-wing and hybrid VTOL drones are useful for large mountainous areas and long infrastructure corridors.

LiDAR payloads are particularly valuable where vegetation or complex terrain is present, while high-resolution RGB cameras provide detailed visual information.

For long-term monitoring, repeatability and georeferencing quality may be more important than maximum camera resolution.

The complete system should therefore be evaluated around the required measurement rather than simply choosing the drone with the largest sensor.

Recommended Multi-Sensor Approach

A strong landslide-monitoring programme may combine several technologies rather than relying on a single payload.

RGB cameras provide detailed visual documentation and photogrammetry. LiDAR provides precise three-dimensional geometry and improved terrain mapping beneath some vegetation. Thermal cameras may help identify certain moisture or drainage patterns. Multispectral cameras can monitor vegetation and surface differences.

These airborne measurements can then be combined with GNSS monitoring, rainfall data, piezometers, inclinometers, ground-based radar, geological mapping and satellite observations.

Each technology contributes a different part of the overall picture.

Future of Drone Landslide Detection

Landslide monitoring is likely to become increasingly automated.

Fixed sensors may continuously monitor rainfall, groundwater and movement. When predetermined thresholds are exceeded, an automated drone could be deployed to collect updated LiDAR and imagery.

AI could compare the new dataset with the previous survey and highlight areas showing significant change.

Satellite monitoring could identify regional deformation, while drones investigate specific locations at much higher resolution.

Digital twins could combine every data source into one continuously updated environment.

Rather than waiting for an engineer to manually compare individual surveys, future systems could automatically identify changes requiring professional attention.

A future monitoring workflow could operate as:

baseline geological and terrain assessment → high-resolution drone LiDAR and photogrammetry survey → permanent monitoring sensors established → rainfall, groundwater and movement monitoring → threshold or scheduled drone deployment → repeat LiDAR/RGB/multispectral collection → precise georeferencing → automated 3D change detection → AI-assisted anomaly screening → comparison with GNSS, piezometer, radar and weather data → geotechnical professional review → targeted ground investigation → updated hazard assessment → continued monitoring or mitigation.

Conclusion

Drones have become an important tool for landslide detection, mapping and long-term slope monitoring because they can collect detailed information across areas that may be difficult or dangerous to access from the ground.

High-resolution RGB imagery can document cracks, erosion and visible changes. Photogrammetry can create detailed three-dimensional models. LiDAR can measure complex terrain and provide improved ground information beneath some vegetation. Thermal and multispectral sensors can add supporting environmental information.

The greatest value comes from repeated measurement.

A baseline drone survey establishes the initial condition of a slope. Subsequent flights can then reveal changes in terrain, cracks, scarps, debris and other surface features. When these measurements are combined with rainfall, groundwater, GNSS, radar and geotechnical instrumentation, specialists gain a much broader understanding of how the slope is changing.

However, surface change does not automatically mean imminent failure, a visible crack does not reveal the complete failure mechanism, a thermal anomaly does not prove groundwater instability, and non-detection of movement does not establish that a slope is safe.

Drone landslide monitoring should therefore complement rather than replace professional geotechnical investigation.

The strongest approach combines drones, LiDAR, photogrammetry, precise positioning, repeat surveys, automated change detection, fixed ground sensors, satellite monitoring, GIS and professional geotechnical interpretation.

As autonomous drones, AI and remote monitoring systems continue to develop, drones are likely to become an increasingly important component of landslide early-warning and infrastructure-monitoring programmes, providing engineers and authorities with faster, safer and more detailed information about changing terrain.