Drone Guide Ship Cargo Hold Inspection

Ship cargo holds are large, complex spaces that require regular inspection to assess structural condition, cleanliness, damage, corrosion and readiness for cargo operations. Traditionally, inspecting these areas can involve personnel entering the hold, working at height, using scaffolding, rope access or mobile access equipment to reach upper structural components. Drones provide an alternative method for conducting an initial visual and sensor-based assessment while reducing the amount of time inspectors need to spend in difficult-to-access areas.

Cargo-hold inspection drones can capture high-resolution imagery, video, thermal data and, with appropriate platforms, three-dimensional measurements of internal structures. They can examine bulkheads, frames, stiffeners, brackets, hatch covers, tank tops and other visible areas from close range. Protective cages, collision-tolerant designs, powerful lighting and GNSS-denied navigation technologies are particularly valuable because cargo holds are enclosed metallic environments where satellite navigation is generally unavailable.

The strongest application is not replacing marine surveyors or classification professionals, but providing them with better access to visual information. A drone can rapidly identify areas requiring closer investigation so that subsequent physical inspection, ultrasonic thickness measurement or other non-destructive testing can be concentrated where it is most valuable.

Why Use Drones Inside Ship Cargo Holds?

Cargo holds can be extremely large. On bulk carriers in particular, structural members may extend many metres above the tank top, making close visual inspection of upper areas difficult without specialist access equipment.

A drone can fly from the lower hold to upper frames, brackets and deck structures within minutes. Cameras can provide close-up imagery that inspectors can review from outside the immediate inspection area or from an appropriate safe position.

This can significantly change the inspection workflow. Rather than providing physical access to every elevated location simply to determine whether closer examination is necessary, the drone can perform an initial visual survey. Areas showing possible corrosion, coating deterioration, cracking, deformation or other anomalies can then be selected for targeted follow-up.

However, drone imagery does not automatically establish structural condition. A visible anomaly may indicate a potential problem, but its significance should be determined by an appropriately qualified marine professional.

Types of Ships

Cargo-hold drone inspection can potentially be used across bulk carriers, general cargo ships and other vessels containing sufficiently large internal cargo spaces. The precise inspection method depends heavily on the vessel design.

Bulk carriers are particularly relevant because of the size and structural complexity of their holds. Frames, brackets, bulkheads and other structures may be difficult to access from the tank top.

General cargo vessels may contain decks, beams and structural arrangements requiring different flight paths. Some cargo spaces may be relatively confined, while others provide substantial room for flight.

A drone system should therefore be selected according to the vessel rather than assuming that one aircraft configuration will work effectively inside every cargo hold.

Pre-Inspection Planning

A successful inspection begins before the drone enters the hold. Available drawings, previous inspection reports and known areas of concern can help determine where detailed imagery is required.

The operator should understand the geometry of the hold, potential obstacles and communication limitations. Lighting, ventilation, dust and remaining cargo material should also be considered.

Flight planning inside a ship is fundamentally different from ordinary outdoor drone surveying. There may be no GNSS, magnetic conditions may be challenging, communications may be partially obstructed by steel structures and the aircraft may be operating close to walls.

A structured inspection plan helps ensure that coverage is systematic rather than simply flying around the space and collecting unstructured video.

GNSS-Denied Flight

GNSS signals are generally unavailable or unreliable inside a steel cargo hold.

The drone therefore needs another method of maintaining stable flight and estimating its position. Depending on the platform, this may involve visual-inertial odometry, optical flow, LiDAR, SLAM or combinations of these technologies.

SLAM is particularly valuable because the drone can use surrounding structural geometry to estimate its movement while simultaneously creating a map of the environment.

However, GNSS-denied navigation should not be confused with guaranteed positioning. Repetitive steel structures and low-light conditions can challenge some localisation systems. The aircraft should therefore be designed specifically for indoor or confined-space operations.

Collision-Tolerant Drones

Collision-tolerant drones are particularly suitable for cargo-hold inspection.

These aircraft typically surround their propellers with a protective cage or frame. If the drone makes light contact with a wall or structural member, the protective structure can reduce the likelihood of propeller damage.

Some platforms can maintain controlled flight after minor contact.

This is valuable inside cargo holds where frames, brackets and other structural elements create complex obstacles.

However, collision tolerance should not be interpreted as permission to deliberately strike structures. Controlled stand-off and careful flight remain important for both aircraft safety and inspection quality.

Lighting

Cargo holds can be extremely dark, particularly away from open hatch covers.

Powerful onboard lighting is therefore essential for conventional visual inspection.

Lighting should illuminate the inspection area without producing excessive glare from painted or metallic surfaces. The relationship between camera, lights and structure is important because poor lighting angles can hide surface features.

Adjustable or directional lighting can help reveal surface texture.

Multiple lights positioned around the camera can reduce shadows, although some shadow can also be useful for highlighting deformation and surface irregularities.

Image quality should be evaluated as an inspection requirement rather than simply maximising brightness.

RGB Cameras

High-resolution RGB cameras are the primary sensor for many cargo-hold drone inspections.

They allow surveyors to examine coatings, corrosion, deformation and other visible features.

Still images are often more useful for detailed analysis than continuous video because individual frames can be examined closely and associated with specific locations.

Video remains valuable for maintaining spatial context.

The strongest workflow can therefore collect both systematic high-resolution photographs and continuous contextual video.

Camera stabilisation, focus and exposure become particularly important when operating close to structural surfaces.

Structural Members

Cargo holds contain numerous structural elements that may require examination.

These can include transverse frames, longitudinal structures, stiffeners, brackets, bulkheads, hopper structures, topside structures and deck-related components.

A drone can approach elevated structures without requiring an inspector to physically access every location.

This provides an efficient way to identify visible anomalies.

However, imagery should be collected from angles that allow the inspector to understand the geometry.

A single image may hide a defect behind a flange or structural member. Multiple viewing angles can provide better evidence.

Corrosion Inspection

Corrosion is one of the most important conditions assessed during cargo-hold inspection.

Drone imagery can identify visible rusting, coating breakdown, scaling and areas where corrosion appears more severe.

The aircraft can capture detailed imagery of areas that would otherwise require working-at-height access.

However, visual corrosion does not directly determine remaining steel thickness.

A surface may appear heavily corroded while retaining significant material, while other areas may contain important thickness loss that is difficult to quantify visually.

Drone inspection should therefore be used to identify candidate areas for closer investigation. Ultrasonic thickness measurement or other appropriate NDT methods remain necessary where quantitative material assessment is required.

Coating Condition

Protective coatings play an important role in reducing corrosion.

Drones can document areas where coatings appear intact, damaged, blistered, cracked or missing.

Repeated inspections can show how coating condition changes over time.

High-resolution imagery can also support maintenance planning by showing the approximate distribution of deterioration.

However, imagery alone may not determine why a coating has failed.

Surface preparation, contamination, mechanical damage and environmental exposure can all contribute.

The drone provides visual evidence that supports subsequent professional assessment.

Cracks and Structural Damage

High-resolution cameras may identify visible cracking or indications that justify closer examination.

The drone can also document buckling, deformation and impact damage.

However, the ability to detect a crack depends on its size, lighting, camera resolution, viewing angle, distance and surface condition.

A non-detection therefore does not prove that no crack exists.

Where imagery indicates a potential crack, inspectors may require closer physical examination or specialist NDT.

Drone imagery should support the inspection process rather than be interpreted as proof that a structure is defect-free.

Deformation

Cargo operations can expose internal structures to significant mechanical forces.

Frames, brackets or other components may become bent or deformed.

Oblique drone imagery can help reveal changes in structural shape.

LiDAR or photogrammetric modelling may provide additional geometric information where quantitative assessment is required.

However, apparent deformation in a photograph can also result from lens perspective.

Engineering conclusions should therefore be based on appropriate measurements rather than visual appearance alone.

Bulkhead Inspection

Bulkheads are large surfaces that can be difficult to inspect comprehensively from the bottom of a cargo hold.

A drone can fly systematic vertical and horizontal passes across the surface.

This provides detailed coverage of weld areas, stiffeners and other visible features.

Structured flight paths make later review easier because imagery can be associated with specific sections.

For repeat inspections, similar routes can be flown again to compare condition over time.

This creates the foundation for a digital inspection record rather than a collection of unrelated photographs.

Tank Top Inspection

The tank top is normally more accessible than elevated structures, but drone imagery can still provide useful documentation.

It may reveal visible coating damage, deformation, corrosion or remnants from previous cargoes.

However, the lower operating height means manual inspection may sometimes provide better detail.

The drone’s main advantage is therefore often the ability to connect tank-top observations with the wider visual record of the complete hold.

Hatch Cover Inspection

Cargo-hold inspections may also include visible parts of hatch covers and surrounding structures.

Drones can capture the underside of covers, beams and difficult-to-access components.

Corrosion, coating deterioration and visible deformation can be documented.

However, visual drone inspection does not by itself confirm hatch-cover weathertightness.

Dedicated testing methods may still be required.

The drone provides complementary visual information.

Cleanliness Inspection

Cargo-hold cleanliness can be important before loading certain cargoes.

Drones can provide an overview of surfaces and identify visible residues or areas that appear to require additional cleaning.

They can be particularly useful for checking elevated ledges and structural areas where material may remain.

However, camera imagery cannot necessarily detect all contamination.

Residues may be chemically present without being visually obvious.

Where cleanliness requirements demand physical testing, sampling or certification, those procedures remain necessary.

Cargo Residue

Material from previous cargoes may accumulate on structural ledges or difficult-to-reach areas.

Drone cameras can help locate visible deposits.

This allows cleaning crews to target specific areas.

After cleaning, another flight can document the result.

However, absence of visible residue does not automatically confirm that the hold meets the requirements for the next cargo.

Professional inspection and any required testing remain important.

Thermal Inspection

Thermal cameras can be integrated with cargo-hold inspection drones.

They measure surface temperature differences rather than directly seeing structural defects.

Thermal information may help identify unusual temperature patterns associated with adjacent machinery, moisture-related effects or other conditions.

However, a thermal anomaly does not identify its cause.

Surface emissivity, airflow and temperature equilibration all affect thermal readings.

Thermal imaging should therefore complement visual inspection rather than be used as a standalone structural assessment.

LiDAR Inspection

LiDAR can create a three-dimensional representation of the cargo hold.

This provides useful spatial context for inspection imagery.

Structural geometry can be documented, and the resulting point cloud can support measurements and digital-twin applications.

SLAM LiDAR is particularly relevant because GNSS is unavailable inside the hold.

However, a LiDAR point cloud does not automatically determine structural integrity.

It measures visible geometry.

Internal corrosion, material thickness and microscopic cracking require other inspection techniques.

3D Cargo-Hold Mapping

A drone equipped with LiDAR or suitable imaging systems can create a three-dimensional model of the cargo hold.

Inspection photographs can then be linked to their approximate locations within this model.

Instead of reviewing hundreds of disconnected images, a surveyor could navigate through a virtual representation of the hold and select particular structures.

This approach can significantly improve inspection documentation.

It also creates a baseline against which future surveys can be compared.

Photogrammetry

Photogrammetry can create three-dimensional models from overlapping photographs.

Inside a cargo hold, this requires good lighting and sufficient visual texture.

Large areas of similar painted steel can sometimes make image matching difficult.

LiDAR may therefore provide more robust geometry in some environments.

However, photogrammetry provides highly detailed colour information.

Combining LiDAR geometry with RGB imagery can produce particularly useful inspection models.

Ultrasonic Thickness Measurement

Ultrasonic thickness measurement is an important technique for determining remaining steel thickness.

Traditional visual drone inspection cannot provide this measurement.

Specialist robotic systems may carry ultrasonic probes and physically contact the structure.

This requires significantly more sophisticated control because the probe needs correct orientation, contact pressure and, depending on the system, appropriate coupling.

The drone may first perform a visual inspection and identify candidate areas. Targeted thickness measurements can then be taken using appropriate equipment.

This can create a more efficient inspection workflow.

NDT Integration

Future cargo-hold drones are likely to combine visual inspection with multiple NDT technologies.

Potential systems could integrate ultrasonic measurements, thermal sensors or other specialist inspection equipment.

However, each technique measures a different property.

LiDAR measures geometry. RGB cameras show visible condition. Thermal sensors measure surface temperature. Ultrasonic systems can measure material thickness under appropriate conditions.

Combining these sensors provides more information, but professional interpretation remains necessary.

Mapping Inspection Findings

One of the major advantages of digital drone inspection is the ability to associate observations with locations.

Instead of reporting simply that corrosion was found somewhere on a bulkhead, inspection software can link an image or observation with a specific structural area.

A 3D model can make this even more intuitive.

This improves communication between shipowners, surveyors, maintenance teams and repair contractors.

Location consistency is particularly valuable when comparing successive inspections.

AI-Assisted Defect Detection

Artificial intelligence can assist with reviewing the large number of images generated during cargo-hold inspections.

Computer-vision systems may highlight candidate corrosion, coating damage or other visible anomalies.

This can reduce the amount of routine imagery requiring immediate manual attention.

However, AI detection should be treated as screening.

A highlighted region is not automatically a structural defect, and an area not highlighted by the algorithm should not automatically be considered defect-free.

Qualified marine professionals should remain responsible for interpretation.

Corrosion Classification with AI

AI systems can potentially categorise visible surface condition and compare corrosion patterns across large areas.

Repeated surveys may allow deterioration to be tracked.

However, appearance varies with lighting, coating colour, rust staining and camera exposure.

Training data therefore strongly influences performance.

AI is most useful for identifying candidate areas and prioritising review rather than independently determining structural condition.

Change Detection

Repeat drone inspections create opportunities for change detection.

Images or 3D models from different inspection dates can be compared.

Software may highlight areas where geometry or surface appearance has changed.

This can support condition monitoring.

However, different lighting and camera angles can create apparent visual changes.

Geometric differences can also arise from mapping uncertainty.

Professional review should therefore determine whether detected change represents genuine deterioration.

Digital Twins

Cargo-hold inspection data can contribute to a digital twin of the vessel.

LiDAR provides geometry, RGB imagery provides visual condition and inspection records provide historical information.

An inspector could navigate through the virtual hold and view observations associated with individual structural components.

Future inspections could update the model.

This would create a long-term record of how visible condition changes over the vessel’s operational life.

However, the digital twin should clearly identify when each dataset was collected.

Inspection Repeatability

Repeatability is important for long-term monitoring.

Using a consistent flight pattern allows similar structures to be photographed during successive inspections.

Autonomous or semi-autonomous flight can improve this consistency.

The drone might follow a planned route around each frame or bulkhead.

However, changes in cargo residue, lighting and vessel condition may require route adjustments.

Automation should therefore support the inspector rather than rigidly prevent adaptation.

Autonomous Cargo-Hold Inspection

GNSS-denied navigation, SLAM and obstacle avoidance are making increasingly autonomous indoor inspections possible.

A drone could map the hold and automatically follow a structured inspection route.

It could identify areas that have not received sufficient image coverage and collect additional photographs.

Future systems may automatically compare new imagery with previous inspection data.

However, autonomous navigation and autonomous inspection interpretation are different capabilities.

A drone may be able to fly itself while still requiring a marine professional to determine the significance of its observations.

Drone-in-a-Box for Maritime Inspection

Although Drone-in-a-Box technology is usually associated with outdoor infrastructure, automated docking systems could eventually support routine vessel inspection.

A vessel might carry a dedicated inspection drone that can be deployed when the cargo hold is empty.

The system could perform repeat routes and upload imagery to a shore-based inspection platform.

This could provide more frequent condition monitoring between formal surveys.

However, maritime movement, salt, humidity and maintenance requirements make onboard automation challenging.

Vessel Movement

A ship may move even when berthed or anchored.

Inside the cargo hold, the drone moves relative to the vessel rather than the earth.

GNSS-independent navigation based on the internal structure can therefore be advantageous.

However, significant vessel motion can affect flight stability.

Cargo-hold drone operations should consider sea state and vessel movement.

Conditions suitable for a conventional visual inspection may not automatically be suitable for stable drone flight.

Magnetic Interference

Steel structures can influence magnetic compasses.

A drone that relies heavily on magnetometer heading may therefore experience navigation difficulties.

Purpose-built indoor platforms often rely more heavily on visual, inertial or LiDAR-based localisation.

The navigation architecture should be understood before operating inside large steel structures.

Compass behaviour experienced outdoors should not be assumed to apply inside a cargo hold.

Communications

Steel bulkheads can interfere with radio communication.

The operator may therefore experience reduced range or temporary signal obstruction.

The drone should have appropriate behaviour if communications deteriorate.

Some platforms can maintain position, retrace their route or continue autonomous navigation.

The correct strategy depends on the system and environment.

Communication coverage should be assessed as part of pre-inspection planning.

Battery Management

Cargo-hold inspection can involve repeated climbing, hovering and close manoeuvring.

These operations consume significant energy.

Protective cages, powerful lighting and specialist sensors add weight.

Battery endurance may therefore be lower than during ordinary outdoor flight.

Large holds should be divided into logical inspection sections.

It is generally preferable to change batteries safely between sections rather than attempting to complete an oversized inspection close to battery limits.

Dust

Bulk cargo operations can leave substantial dust inside holds.

Dust can affect camera lenses, lighting and LiDAR sensors.

Rotor downwash may also disturb settled material.

This can reduce visibility.

The inspection should therefore ideally take place after appropriate cleaning and settling time.

Sensor windows should be checked regularly.

Heavy airborne dust may make useful inspection impossible until conditions improve.

Moisture and Condensation

Cargo holds can contain moisture and condensation.

This can affect visibility and create reflections.

Water droplets on camera or LiDAR windows can degrade data quality.

Moisture can also influence the appearance of surfaces and thermal patterns.

Environmental conditions should therefore be recorded.

An observation made on a wet surface may not be directly comparable with imagery collected when the structure is dry.

Hazardous Atmospheres

Some cargoes or vessel conditions can create hazardous atmospheres.

A conventional drone should not be assumed suitable for an explosive or otherwise hazardous environment.

Gas testing, ventilation and vessel safety procedures may be required before entry.

If an environment requires intrinsically safe equipment, the drone and all associated electronics must meet the applicable requirements.

The use of a drone does not remove confined-space or hazardous-atmosphere responsibilities.

Confined-Space Safety

A cargo hold may be treated as an enclosed or confined space depending on the vessel, operation and applicable procedures.

Using a drone can reduce the need for personnel to enter certain areas, but it does not automatically remove the safety controls associated with the space.

The vessel’s safety management procedures remain central.

Drone operations should be coordinated with the responsible ship personnel and inspection team.

Human Entry Reduction

One of the strongest benefits of drone inspection is reducing unnecessary human exposure.

An initial drone survey can identify where close physical access is actually required.

Instead of arranging access to every upper frame, the inspection team can concentrate on selected locations.

This can reduce working-at-height exposure and potentially shorten inspection preparation.

However, where regulations or professional judgement require physical access, the drone should not be used as a reason to avoid it.

Working at Height

Traditional inspection of upper cargo-hold structures can require scaffolding, cherry pickers, rope access or rafting depending on the vessel and circumstances.

Drones can capture visual information without placing a person at height.

This is one of their clearest safety advantages.

The benefit is particularly strong during preliminary condition assessment.

If the drone identifies an area requiring physical measurement or repair, appropriate access can then be provided specifically for that location.

Inspection Time

Drone inspections can potentially reduce the time required to visually access large cargo holds.

A drone can move rapidly between upper structures.

Digital imagery can also be reviewed after collection.

However, faster data acquisition does not automatically mean a faster complete inspection.

Large volumes of imagery require organisation and review.

Structured data collection and AI-assisted screening can help prevent inspection teams from simply replacing access time with excessive data-review time.

Image Traceability

Every inspection image should ideally be associated with a known location.

Without traceability, thousands of photographs can become difficult to use.

A structured naming system, flight route or 3D model can solve this problem.

Inspection software can link images to frame numbers, bulkheads or structural components.

This makes findings easier to communicate and supports repeat inspection.

Data organisation should therefore be planned before the flight rather than after hundreds of images have already been collected.

Reporting

Drone inspection reports should distinguish observations from conclusions.

For example, the report may identify an area showing visible coating breakdown and corrosion-like surface appearance.

A qualified surveyor can then determine whether further investigation is necessary.

This is more appropriate than automatically describing every visual anomaly as a structural defect.

Reports should also identify areas that could not be adequately inspected.

Incomplete coverage should be transparent.

Classification Society and Regulatory Requirements

Drone inspection can support maritime survey and inspection workflows, but acceptance depends on the applicable vessel, flag-state, classification and inspection requirements.

The use of a drone does not automatically replace required close-up surveys, thickness measurements or other mandated inspection activities.

Operators should therefore confirm the requirements applicable to the specific inspection.

The most effective model is often to integrate drone inspection into an approved professional survey process rather than treating it as an independent replacement.

Data Security

Cargo-hold imagery and vessel models can contain commercially or operationally sensitive information.

Inspection data should therefore be handled appropriately.

Access permissions, cloud storage, encryption and retention policies may be important.

Where third-party processing platforms are used, shipowners should understand where data is stored and who can access it.

Detailed 3D vessel models may warrant stronger controls than ordinary inspection photographs.

Selecting a Drone for Cargo-Hold Inspection

The best cargo-hold inspection drone is not necessarily the aircraft with the longest flight time or highest-resolution camera. The complete system should be designed for enclosed, GNSS-denied and obstacle-rich environments.

Important considerations include collision protection, low-light camera performance, onboard lighting, GNSS-independent localisation, LiDAR or visual SLAM, obstacle awareness, communications resilience, flight stability, payload capability and battery endurance.

Aircraft size also matters. A smaller drone may access structural spaces more easily, while a larger platform may carry better sensors.

The correct balance depends on the vessel and inspection requirement.

Selecting Inspection Payloads

A basic inspection system may require only an RGB camera and lighting.

More advanced operations may integrate thermal imaging or LiDAR.

Specialist platforms may carry contact-based NDT sensors.

Payload selection should therefore begin with the inspection question.

If the requirement is to identify visible coating deterioration, high-quality RGB imagery may be sufficient.

If the requirement is to measure geometry, LiDAR may be appropriate.

If the requirement is remaining steel thickness, an appropriate ultrasonic or other approved measurement technique is required.

Using the correct sensor is more important than carrying the largest number of sensors.

Benefits and Limitations

Drone cargo-hold inspection can provide major benefits in access, documentation, efficiency and personnel safety. It is particularly valuable for elevated structures and locations that would otherwise require substantial access equipment simply to perform an initial visual examination.

Drones can capture high-resolution imagery, video, thermal information and 3D geometry, while digital systems can associate observations with precise areas of the hold. Repeat inspections can create a long-term record of visible condition.

However, there are important limitations. A camera cannot directly determine steel thickness. LiDAR cannot reveal internal defects. Thermal anomalies do not automatically identify structural problems. AI detection cannot confirm that a structure is safe, and an area that appears normal in drone imagery may still require physical inspection.

A drone therefore provides inspection evidence rather than independent engineering judgement.

The Future of Drone Cargo-Hold Inspection

Cargo-hold inspection is likely to become increasingly digital and automated.

Future drones may autonomously navigate complete holds using LiDAR and visual SLAM, following repeatable routes around frames, bulkheads and deck structures. High-resolution imagery could automatically be associated with individual structural components within a 3D vessel model.

AI could compare the latest inspection with previous surveys and highlight candidate areas showing increased corrosion, coating deterioration or geometric change. These observations could then be prioritised for marine surveyor review.

Specialist robotic drones may increasingly perform contact measurements. A drone could conduct the initial visual survey, identify candidate corrosion areas and then position an ultrasonic probe against selected surfaces to collect thickness measurements.

Integration with vessel digital twins could create a long-term structural inspection record in which photographs, measurements, repair history and survey observations are associated with their physical locations.

A future workflow could operate as:

inspection requirement → vessel drawings and previous records reviewed → cargo hold prepared and safety conditions verified → GNSS-denied drone deployed → systematic RGB/LiDAR inspection → AI-assisted candidate anomaly identification → observations mapped onto 3D cargo-hold model → marine surveyor review → targeted close-up inspection or NDT where required → maintenance or repair → follow-up drone documentation → vessel inspection record and digital twin updated.

Conclusion

Drones are becoming a valuable tool for inspecting ship cargo holds because they can rapidly access elevated and difficult-to-reach structures while reducing the amount of work that must initially be performed at height.

Collision-tolerant aircraft equipped with high-resolution cameras, powerful lighting and GNSS-independent navigation can inspect frames, bulkheads, brackets, hatch-cover structures and other visible areas. More advanced systems can integrate LiDAR, thermal imaging and specialist NDT technologies.

Their greatest value lies in creating a safer and more efficient inspection workflow. Instead of providing physical access to every structure simply to determine its condition, a drone can perform an initial digital survey and help identify where closer investigation is justified.

However, drone imagery should not be confused with structural certification. Visible corrosion does not directly determine remaining steel thickness, a thermal anomaly does not establish its cause, a LiDAR model does not reveal internal defects, and non-detection does not prove that a structure is defect-free.

The strongest programmes therefore combine drone inspection with qualified marine surveyors, appropriate NDT methods, structured reporting and targeted physical verification.

As autonomous navigation, SLAM, AI-assisted inspection and robotic NDT continue to improve, drones are likely to become increasingly integrated into ship inspection programmes, providing shipowners, operators, surveyors and maintenance teams with faster, safer and more repeatable information about the condition of cargo-hold structures.