Reducing Dam Maintenance Labor: 6 Signs of Deterioration Identified from Drone Point Clouds

Table of Contents

• Efficient detection and recording of cracks

• Three-dimensional understanding of delamination and missing parts

• Quantitative monitoring of dam deformation

• Wide-area visualization of leakage stains and discoloration

• Remote confirmation of vegetation overgrowth on dam surfaces

• Safe inspection of foundation and surrounding erosion

• Conclusion

In recent years, dam aging has been progressing in Japan. There are thousands of dams across the country, many of which were built around the period of rapid economic growth and have been in service for more than half a century. Regular inspections and appropriate maintenance are indispensable to maintain safe water supply and disaster prevention functions. However, dam inspections involve large areas and high-elevation work, requiring significant labor and time and posing safety risks. Traditionally, workers approached dam faces by erecting scaffolding or descending on ropes to perform visual inspections and sounding tests. These methods not only require great effort but also restrict the inspection area and carry the risk of overlooking abnormalities.

To address these challenges, inspections using drones have attracted attention in recent years. By equipping aerial drones with high-performance cameras and laser scanners and capturing aerial and close-up images of the entire dam to obtain three-dimensional point cloud data, precise assessment from a distance has become possible. Drone inspections can safely access dangerous locations that humans cannot enter and cover wide areas in a short time, greatly contributing to labor savings and improved safety. Analyzing acquired point cloud data also enables detection of fine deterioration signs that were hard to find conventionally and allows quantitative understanding of long-term changes. As a result, early repair planning and preventive maintenance in dam management become feasible, helping to prevent major accidents.

With initiatives such as the Ministry of Land, Infrastructure, Transport and Tourism’s i-Construction and infrastructure management DX, the adoption of drones and 3D data technology for dam inspections is accelerating nationwide. In actual dam management, advanced efforts have begun, such as using AI to analyze drone images to automatically detect cracks and trialing systems that enable remote real-time monitoring, aiming to alleviate labor shortages and improve inspection accuracy. Smart maintenance leveraging the latest technologies will become increasingly important to ensure infrastructure safety with limited human resources.

This article carefully selects and introduces six deterioration signs of dams that can be identified using point cloud data acquired by drones. Efficient detection and monitoring of these signs can be expected to reduce inspection labor and improve reliability. Each symptom is explained in detail regarding how it can be identified from point clouds and how it is useful in maintenance practice.

1. Efficient detection and recording of cracks

“Cracks” that appear on concrete dam surfaces are one of the representative signs of deterioration. Even fine cracks, if left untreated, allow water ingress that accelerates structural degradation, so early detection and recording are important. Traditionally, inspections relied on visual checks and crack gauges for width measurement, but working at height is difficult and human visual inspection has limitations.

Combining high-resolution images captured by drones with point cloud data enables exhaustive detection of cracks across the vast dam surface. By mapping images onto the point cloud, crack locations can be accurately recorded in three-dimensional coordinates, and their length and width can be measured digitally. For example, even fine cracks on the order of several millimeters in width can be captured without omission by using high-detail images taken from close range. Drones can safely approach locations where people cannot, allowing comprehensive surveys of the entire dam and greatly reducing the risk of overlooking cracks. Furthermore, the distribution and orientation of detected cracks help infer stress patterns and degradation causes affecting the dam. Visualizing these on a 3D model makes it easier to evaluate the behavior of the entire structure comprehensively.

If the acquired point cloud data and photographs are stored, comparing them with data from the next inspection makes it possible to understand crack progression. Regular 3D monitoring enables quantitative detection of changes such as “a crack has become longer than before” or “a new crack has appeared,” allowing early recognition of deterioration trends. Recently, technologies that automatically extract cracks by AI analysis of captured images and map them onto point cloud models have emerged, reducing the burden of manual interpretation. Utilizing drone point clouds dramatically streamlines crack detection and recording and aids in repair planning. By referring to crack information mapped on the point cloud model, it is possible to pre-evaluate which areas should be prioritized for repair and allocate limited resources efficiently. As a result, continuous monitoring of dam health becomes easier.

2. Three-dimensional understanding of delamination and missing parts

Delamination, where the concrete surface peels off, and spalling, where chunks fall off, are also frequently observed deterioration phenomena in dams. When the concrete surface degrades due to freeze–thaw cycles from temperature differences or long-term weathering, concrete pieces become prone to detachment, eventually creating missing areas. Such delamination and spalling not only reduce the durability of the structure itself but also pose the danger of detached concrete pieces causing damage to people or equipment, so early detection and countermeasures are required.

Point cloud data obtained by drones accurately reflect subtle irregularities on the dam surface, allowing three-dimensional understanding of dents caused by delamination and missing parts. Small-scale delamination that is difficult to see with the naked eye can be clearly identified on a 3D model as a height difference from intact areas. Combined with camera images, signs of delamination such as surface discoloration or exposed rebar can also be detected. For example, if rainwater infiltrates and corrodes rebar inside the concrete, the expansion pressure can lift surrounding concrete, causing cracks and rust-streak stains; drone inspections can capture such anomalies early.

Using point cloud models, it is also easy to measure the size and depth of cavities resulting from spalling. This enables objective evaluation of damage severity and helps determine repair priorities. Areas that used to be observed from a distance with binoculars or telephoto lenses can now be photographed closely by drones, reducing missed small delaminations and high-elevation losses. A major advantage is that inspectors do not need to climb dangerous heights to safely grasp the state of delamination and spalling. By performing three-dimensional observations with drone point clouds, locations of delamination and missing parts can be accurately identified, enabling timely repairs, installation of fall-prevention nets, and other appropriate measures. Periodic acquisition of point cloud data and comparison with past data also allow tracking of new delamination or spalling occurrences and enlargement of existing defects, providing information for timing repairs.

3. Quantitative monitoring of dam deformation

Slight deformation and displacement of the entire dam progresses gradually due to long-term loading, ground movement, earthquakes, and other factors. In concrete dams, temperature changes or alkali–silica reaction (ASR) can cause the dam body to bulge, while rockfill or earthfill dams can experience settlement or local sliding that lowers or tilts the crest. Early detection of these deformations is extremely important for safe dam management.

Traditionally, deformation monitoring has used surveying with reference points or measurements from instruments embedded inside the dam. However, because measurement points are limited, it has been difficult to understand overall changes across the surface. High-density point clouds acquired by drones allow comprehensive measurement of the dam surface shape, so even very small displacements are not overlooked. For example, overlaying a current point cloud model with a previously obtained model or the as-designed reference model enables visualization of the amount of change in each part as a color map. If the central part of the dam is protruding by several centimeters, that bulge can be detected, and relative displacement with the left and right banks becomes clear. Microscopic distortions that human eyes cannot notice can be quantitatively captured by 3D data analysis. In recent years, advances in high-precision sensors and surveying technology mounted on drones have enabled attempts to capture micro-displacements at the millimeter level.

Note that planar deformation data obtained by drones can further improve monitoring accuracy when combined with conventional point-based instruments installed on dams, such as inclinometers and extensometers. Complementing continuous sensor measurements with drone point cloud overview data can build a robust monitoring system that does not miss signs of abnormalities.

Such quantitative monitoring enables early detection of abnormal signs in dam bodies and supports measures such as reinforcement and increased surveillance as needed. Capturing precursors before large-scale deformation occurs provides time to consider disaster-reduction measures in advance. Data obtained from point clouds also serve as objective evidence in inspection reports. Deformation monitoring using drone point clouds greatly contributes to long-term health assessment and risk management of dams.

4. Wide-area visualization of leakage stains and discoloration

“Stains” and discoloration of concrete surfaces caused by leakage or seepage from dams are also important deterioration indicators. When water travels behind the dam, it can seep out onto the concrete surface, producing dark wet marks or white deposits (efflorescence). These leakage stains suggest internal water paths or the presence of cracks and, if left untreated, can expand leakage and lead to structural weakening or piping failure.

Drone inspections enable viewing the distribution of surface discoloration and stains across the entire dam from high-resolution images captured over a wide area. Fine leakage stains that are easily missed when looking up from below can be readily identified in close-range aerial photos or generated orthophotos. For example, if a long black stain appears near the center of the dam, it is likely that water is seeping out at that location from the upstream side, indicating the need for a detailed investigation of the backside at the same point. If many white crystalline deposits are observed, it indicates moisture evaporating inside the concrete and mineral deposits accumulating, suggesting leakage through invisible microcracks.

By overlaying color-textured images onto a drone point cloud, discoloration caused by leakage can be precisely marked on a 3D model. This makes it easier to identify the spatial relationships and extent of stains and to pinpoint abnormal areas. If necessary, equipping drones with infrared cameras can detect moisture distribution invisible to the naked eye by observing surface temperature variations. Such wide-area, detailed visualization enables early detection of leakage-induced deterioration signs, leading to measures such as considering waterproofing works or strengthening monitoring. Comparing saved discoloration maps as orthophotos over time also allows analysis of whether stains are expanding or shrinking, which is useful for post-repair follow-up.

5. Remote confirmation of vegetation overgrowth on dam surfaces

The sight of weeds and small trees taking root on concrete surfaces is a warning sign that must not be overlooked in dam maintenance. Concrete structures are normally managed to prevent plant growth, but soil and dust that accumulate in cracks or perpetually damp areas can allow weed germination and proliferation. Plant roots can widen cracks in concrete and accelerate deterioration, so early removal is necessary. On earthfill or rockfill dam slopes, growth of large trees, like animal burrows, can be a risk factor for internal penetration and should also be regularly inspected and subject to weeding or cutting. Even small plants rooted on concrete can damage structures, so prompt removal is important upon discovery. If drone inspections can reliably detect such vegetation, weeding and cleaning plans can be executed reliably, contributing to dam health maintenance.

Searching for such vegetation overgrowth across a large dam by manpower alone is not easy. Aerial photography by drone makes it simple to find green-discolored areas or plant shadows in panoramic images or photos of the dam surface. Grass and trees growing on high or steep slopes can be safely assessed remotely. By applying aerial images as textures to point cloud data, vegetation locations can be accurately mapped in three-dimensional space, allowing workers to identify problem areas in advance and prepare for on-site response.

Moreover, locations where vegetation proliferates are often continuously moist or associated with cracks and leakage. Therefore, the presence of vegetation can serve as an indicator of dam health. Regularly recording vegetation distribution across the dam with drone inspections enables quick detection of newly emerging plants in abnormal locations, prompting cause investigation and removal before issues escalate. The ability to check hard-to-reach areas remotely significantly contributes to labor savings and improved safety in inspection work.

6. Safe inspection of foundation and surrounding erosion

Not only the dam body but also the foundation and surrounding ground conditions are critical points in maintenance. Scouring of the foundation rock of concrete dams due to long-term discharge and erosion or collapse on downstream slopes of earthfill dams directly affect dam stability. For example, powerful flows from spillways can scour foundation ground and create cavities, or sediment may be lost at intake attachments on both banks, creating steps. There are reported cases where long-overlooked leakage formed cavities in the ground beneath the dam, eventually requiring large-scale repairs. Furthermore, if the riverbed downstream of the dam is gradually scoured and the foundation becomes exposed, bearing capacity decreases and danger ensues. Such anomalies in surrounding areas should be discovered early and appropriately reinforced or backfilled.

Traditionally, inspecting foundations and downstream areas was difficult and hazardous. It required walking slippery steep slopes for visual checks or approaching by boat for observation, limiting the range that could be thoroughly investigated. Using drones, close-up imaging of difficult-to-access, unsafe areas is possible, allowing detailed recording of terrain around the foundation and exposed bedrock. Point cloud data can quantitatively capture surface irregularities and erosion depth of bedrock, so comparing with previous data makes it possible to evaluate erosion progression. For example, you can numerically grasp, from point cloud differences, how much specific terrain has lowered from last year to this year.

Additionally, areas that are hard to inspect manually—such as lakeside slopes or the downstream channel below spillways—can be monitored over wide areas with drones. Combining drone inspections with underwater drones or acoustic surveys when necessary enables comprehensive inspection that includes underwater foundations and riverbed conditions. In any case, drone technology has made it possible to safely and efficiently understand foundation and surrounding conditions that used to require hazardous work, significantly reducing fall accidents during inspections and risks from sudden releases. This contributes greatly to confirming the overall health of dams.

Conclusion

By utilizing drones and point cloud data, dam maintenance can be greatly streamlined while improving accuracy and safety. Detecting deterioration signs such as cracks and delamination without omission and monitoring them quantitatively on 3D models has made it easier to identify wide-area abnormalities that were conventionally difficult to grasp. The introduction of digital technologies also allows flexible increase of inspection frequency, enabling finer tracking of changes. In fact, demonstrations of drone inspections have already been conducted at some domestic dams, reporting outcomes such as significant reductions in work time and decreases in human error.

Furthermore, combining drone inspections with new positioning technologies further increases the efficiency of data acquisition. For example, using the high-precision GNSS positioning device “LRTK” that can be attached to an iPhone makes it easy to obtain centimeter-level position coordinates (half-inch accuracy) on site. If reference points for important parts are surveyed with LRTK, accurate coordinate references can be provided to point cloud models obtained by drones, facilitating understanding of long-term changes and smooth cross-referencing of multiple datasets. By pursuing efficiency with such advanced technologies, it becomes possible for limited personnel to protect dam safety over the long term.

With further advances in drone and point cloud technologies, maintenance of various infrastructure facilities—not just dams—is expected to become smarter and less labor-intensive. Fortunately, these technologies are becoming easier to handle year by year, and miniaturization and cost reduction of drones and positioning devices are making adoption feasible even for small-to-medium management bodies. Dams are critical infrastructure that support people’s lives and safety. It is essential to continue actively adopting the latest technologies and promote smart maintenance to ensure their safety and functions for the next generation.