Five Items to Verify the Survey Accuracy of Solar Power Plants Using Drone Surveys

In the operation and maintenance of solar power plants, it is important not only to inspect the power generation equipment itself but also to continuously monitor the condition of the entire site. Subsidence of the prepared ground, deformation of slopes, changes in drainage routes, terrain changes around mounting racks, and ruts or washouts on access roads can be overlooked during routine patrols. One effective measure is drone surveying, which can record wide areas from above.

However, drone surveying does not automatically produce accurate results simply by flying. To verify surveying accuracy, it is necessary to align multiple aspects: control points, imaging conditions, validation points, resulting data, and on-site verification. In particular, at solar power plants, reflections from panel surfaces, sloped terrain, weeds, fences, mounting racks, drainage structures, and similar factors can affect survey results, so a management perspective different from that of general aerial photography is required.

In this article, we explain five items that solar power plant operational staff should check when verifying the results of drone surveys, presented in a way that closely reflects on-site operations.

Table of Contents

• Verify the consistency between the reference points and the coordinate system.

• Check that the flight plan and imaging conditions meet the accuracy requirements.

• Check the placement of calibration points and verification points.

• Check for misalignment between point clouds, orthophotos, and 3D deliverables

• Maintain surveying accuracy continuously through field verification and record management

• Summary

Check consistency between reference points and coordinate systems

The first thing to check in drone surveying of solar power plants is which reference is being used to align positions. When people talk about survey accuracy, attention tends to focus on image sharpness and the density of point clouds, but if the overall location of the deliverables is offset, the data will be difficult to use in practice no matter how good it looks. Especially when comparing against existing grading plans, drainage plans, racking layout drawings, access road maps, or previous survey results, it is essential that the coordinate system and the handling of control points are consistent.

In solar power plants, the reference systems used for drawings and survey results during design, construction, and maintenance can differ. The drawings from initial construction may be managed in design coordinates, while post-completion management documents may use a different local coordinate system. Additionally, past survey results may have been created using arbitrary on-site reference points. When a new drone survey is conducted under these conditions, overlaying the results can show misalignment, making it difficult to determine whether the discrepancy is due to actual deformation or simply differences in coordinate settings.

The primary points to check are the locations of the reference points and how they are used. When using reference points, boundary stakes, known points, or fixed control points remaining on site, verify that they are still stable. On a power plant site, points used in the past may have shifted due to post-construction settlement, slope repairs, maintenance road improvements, drainage works, and similar activities. Even if they appear fine visually, if there have been changes to the surrounding paving or ground conditions, careful verification is required before using them as reference points.

Next, clarify the coordinate system of the survey results. It is important to decide in advance whether to use public coordinates, to manage them with a local coordinate system within the power plant, or to match the conditions of past results. For maintenance purposes, it is essential that comparisons can be made under the same coordinate conditions each time. For example, when tracking the locations of poor drainage or the extent of slope deformation over time, changing the reference between surveys can lead to misreading the actual amount of change.

Vertical reference should not be overlooked. At solar power plants, even slight variations in ground surface elevation can be related to drainage issues, muddy conditions, or scour. Even if horizontal positions align, if the vertical reference is not consistent, the data will be difficult to use for checking settlement of the graded surface or drainage slopes. When utilizing elevation data obtained by drone surveying, confirm which vertical reference is being used, whether it can be compared with past data, and whether it is consistent with local control points.

In checking reference points, pre-survey records are also important. Organizing the reference point's name, location, photographs, intended use, coordinate values, installation condition, and surrounding conditions makes it easier to assess the results later. This is especially important at power plants where multiple management companies, surveying firms, and construction contractors are involved, so that results can be handled under the same assumptions even if personnel change. If reference point information depends on verbal communication or an individual's memory, the same conditions cannot be reproduced in the next survey.

When verifying the accuracy of drone surveys, you need to look not only at the deliverables themselves but also at which reference standards those deliverables are tied to. Even if the images and point clouds look good, if the way control points were established is ambiguous, the reliability of the results as management documents decreases. Conversely, if the control points and coordinate system conditions are clearly defined, it becomes easier to use the data for comparing periodic surveys, checking before-and-after repairs, evaluating the effectiveness of drainage improvements, and understanding terrain changes.

In managing solar power plants, it is important not to treat survey results as one-off records but as documents that can be compared continuously. To achieve this, you need to organize reference points and coordinate systems at the initial stage and establish a foundation that enables verification of survey accuracy. Even when aiming to enhance the benefits of introducing drone surveying, this is a basic yet critical check; overlooking it can significantly affect downstream processes.

Verify that the flight plan and shooting conditions meet the required accuracy

The accuracy of drone surveying is heavily influenced by flight planning and imaging conditions. Solar power plants have panels arranged regularly across large sites, so they may appear easy to capture at first glance. However, in reality there are panel surface reflections, racking shadows, sloped terrain, embankments, fences, electrical equipment, access roads, vegetation, and other factors, and if conditions are handled incorrectly the survey results can easily become uneven. To verify surveying accuracy, it is essential to check not only the output data but also the conditions under which the flight was conducted and how the images were captured.

The first thing to confirm is the flight altitude. If the flight altitude is too high, you can capture a wide area in a single image, but it becomes difficult to detect fine changes on the ground. Conversely, if it is too low, image resolution improves, but the number of images increases, and processing load and flight time grow. For surveying solar power plants, it is important to set the altitude according to the management purpose. The required shooting conditions vary depending on whether you want to grasp overall terrain changes, inspect the fine condition of drainage channels and slopes, or check level differences and scouring around the mounting racks.

Next, check the image overlap. In drone surveying, multiple photos are overlaid to create orthomosaics and point clouds, so if front-to-back and side-to-side overlap are insufficient, processing can become less stable. At solar power plants, panels are arranged continuously in the same shape, which can make it difficult to distinguish feature points. In areas with repetitive patterns, if sufficient overlap is not ensured, errors are more likely to occur when stitching images. Especially at large-scale plants or plants with slopes, it is necessary to confirm that overlap is not lacking at the edges of the survey area or in places with changes in elevation.

The direction of imaging is also important. Straight-down imaging is suitable for capturing the shape of the ground surface, but slopes, the sides of structures, and areas around drainage facilities can be difficult to assess from directly above alone. If the purpose of the survey is to confirm topography, straight-down imaging is the standard; however, if it also needs to serve for condition assessment or to record repair extents, combining oblique-angle shots as necessary makes it easier to understand site conditions. However, it is important to manage separately the images that will be processed as survey deliverables and the auxiliary images used for visual inspection. If purposes are mixed, it becomes difficult to tell which images were used for accuracy verification.

In solar power plants, attention must also be paid to the time of day when imaging. Solar panels easily reflect light, and the appearance of images changes depending on the time of day and the weather. Strong reflections can cause parts of an image to become overexposed or make feature points difficult to detect. Also, during times when mounting racks or panels cast long shadows, the condition of the ground surface may be hard to see. Shadows and reflections cannot always be completely avoided, but to verify surveying accuracy it is important to record the weather, time of day, and sunlight conditions at the time of imaging.

Do not overlook the effects of wind. Strong winds can make it difficult for the aircraft to maintain a stable attitude, which can affect shooting positions and image quality. Photovoltaic power plants are often installed in open locations and can be particularly susceptible to wind depending on the surrounding terrain and the season. Even if flight itself is possible, whether the quality required for surveying purposes can be ensured is another matter. If you prioritize surveying accuracy, check the flight logs and captured images for significant image blur, irregularities in shooting intervals, or large deviations in the flight path.

Also, defining the survey coverage is an important practical consideration. At solar power plants, the items to be checked may include not only the panel installation area but also drainage channels, retention ponds, slopes, access roads, the fence perimeter, entrances and exits, material storage yards, and surrounding water flows. If the survey area is too narrow, you may not be able to determine the causes of problems occurring within the plant. For example, if muddy patches on the site are related to runoff entering from the perimeter, surveying only around the panels will not reveal the cause. When checking survey accuracy, it is important to look not just at whether the central area is captured cleanly but whether any areas required for management are missing.

Flight planning and imaging conditions also affect the reproducibility of survey results. When conducting drone surveys regularly, large changes in flight conditions from one session to the next make it difficult to compare outcomes. Although it is impossible to keep conditions exactly the same because of season and weather, standardizing flight altitude, coverage area, image overlap, the use of ground control points, and processing parameters as much as possible will make time-series comparisons more reliable.

When reviewing the results of a drone survey, it is important not to judge solely by whether the images look good. To verify surveying accuracy, check whether the flight plan matched the objectives, whether the shooting conditions were reasonable, and whether site-specific influences were taken into account. To produce results usable for managing solar power plants, it is essential to be mindful of accuracy verification from the planning stage prior to capture.

Verify the placement of calibration and verification points

In verifying the accuracy of drone surveying, how control points and check points are handled is important. Control points are used as references to align images and point clouds to real-world coordinates. Check points, on the other hand, are used to confirm how accurately the generated outputs reproduce positions. If these two roles are confused, the outputs may look tidy while objective accuracy verification is insufficient.

At solar power plants, careful planning is required for the placement of control points. Because panels are widely arranged across the site, locations where control points can be placed on the ground surface may be limited. It is necessary to choose locations that are easy to see and stable, such as access roads, vacant areas, between rows of panels, the perimeter, or the edges of graded slopes. However, if control points are concentrated only around the perimeter, verification of accuracy in the central part of the site can be weakened. In large plants, it is desirable not only to arrange points so they surround the entire survey area but also to place appropriate points inside the site.

It is important that calibration points can be clearly identified in the images. If they are hidden by grass, fall into the shadow of panels, or blend into ground patterns, they become difficult to read accurately. At solar power plants, weed growth and the positions of shadows change with the seasons, so locations that were usable in the past may not be usable this time. After installing calibration points, verify that they are visible from above, not at the edge of the imaging area, and not affected by shadows or reflections.

Validation points should be treated separately from control points to verify the accuracy of the results. If you use all known points as control points, you will lack material to externally validate the results. By retaining validation points, you can check how well the produced orthophotos and point clouds align with the field coordinates. When using drone survey results as management documents, the presence or absence of this validation directly affects their reliability.

Ensure that the placement of verification points is not biased. If checks are performed only near the power plant entrances/exits and along access roads, the accuracy on slopes, drainage facilities, the center of the site, and the site edges may be unknown. In particular, if there are areas where you want to check for poor drainage or changes in ground conditions, placing verification points around those areas will make it easier to improve the accuracy of identifying problem locations. However, do not force points into areas that cannot be entered safely or where there is a risk of contacting equipment; instead, consider placements that suit the site conditions.

The measurement methods for control points and check points should also be reviewed. Recording how the point coordinates were obtained, the conditions at the time of measurement, and whether remeasurements or verification measurements were performed makes it easier to evaluate the results. If the accuracy of a drone survey falls short of expectations, the cause can be measurement errors in the control points themselves rather than the flight or image processing. For example, mistakes such as measuring the center of a control point incorrectly, mixing up point names, entering incorrect numbers when inputting coordinates, or the installation location shifting during imaging can affect the entire outcome.

Managing point names is also important at solar power plants. On large sites there are many similar locations, and rows of panels and access roads often look the same, so if point names and positional relationships are ambiguous, mix-ups can easily occur. Assign clear, easy-to-understand names to reference points and verification points, and manage them with layout diagrams, site photographs, and a list of coordinates. If point names and photos do not match, it will cause significant confusion when verifying results later.

In accuracy verification, you should not only look at the errors of the validation points as numbers, but also check how the errors manifest. Whether the offsets shift overall in the same direction, whether only certain areas show large shifts, or whether there is a bias in the vertical direction will change the diagnosis of the cause. If the shift is overall, there may be a problem with the coordinate system or the setting of reference points. If only part of the area is shifted, possible causes include insufficient image overlap, reflections, shadows, lack of surface features, or biased placement of control points.

Also, the results at verification points need to be evaluated together with the intended use of the outputs. The way required accuracy is considered differs between using the data for broad-area situational awareness and using it for purposes closer to serving as the basis for repair quantities or for detailed as-built confirmation. In the maintenance and management of solar power plants, rather than demanding the same accuracy for all uses, it is practical to organize the necessary level of reliability for each purpose, such as drainage checks, monitoring terrain changes, repair planning, and documentation.

Ground control points and check points are mechanisms for giving objectivity to the results of drone surveying. Even if the outputs look visually clean, if they haven't been verified at check points it is difficult to judge how much they can be trusted. When using drone surveying to manage a solar power plant, it is important to verify the entire workflow—from point placement and measurement to naming management and the recording of verification results.

Check for misalignment between point clouds, orthophotos, and 3D outputs

Typical deliverables produced by drone surveying include orthomosaic images, point clouds, 3D models, elevation data, and cross-section data. In the management of solar power plants, these are used to verify the current conditions across the entire site, terrain changes, drainage routes, slope conditions, deformation or damage to access roads, and subsidence around mounting racks. However, because much of the deliverable data is generated by automated processing, it is necessary not to accept the finished data at face value but to check for discrepancies and any unnatural areas against on-site conditions.

The first thing to check is the positional shift of the orthoimage. Because an orthoimage can be treated like a map viewed from above, it is often used overlaid with past drawings and surveying results. What you should pay attention to here is whether the positions of panel rows, management roads, drainage ditches, fences, buildings, and so on, are consistent with existing materials. If the entire image is shifted uniformly in one direction, the cause may lie in the coordinate settings or the handling of control points. On the other hand, if only parts appear stretched or bent, you should suspect influences from image processing or shooting conditions.

In solar power plants, reflections from panel surfaces and the repeated arrangement of identical shapes can cause image processing to become unstable. In areas where similar patterns repeat, matching points between images can be incorrectly identified. As a result, parts of the orthophoto may become distorted and the point cloud may show unnatural bulging. In particular, if panel surfaces are treated as the ground surface, this can lead to misunderstandings when you want to check ground elevation. When reviewing survey results, it is important to distinguish panels and mounting structures from the ground surface.

With point cloud data, check how accurately the ground surface has been captured. In solar power plants, the ground beneath panels may be difficult to capture in images. Also, in areas where weeds are dense, the tops of the vegetation may be captured as point cloud and appear higher than the actual ground. If you want to check drainage gradients or settlement, treating vegetation-affected point cloud as the ground can lead to incorrect assessments. As needed, operational measures such as surveying after mowing, choosing a season when the ground surface is visible, and combining with on-site verification are effective.

Vertical discrepancies are a point to pay particular attention to in the operation and maintenance of solar power plants. Poor drainage within a plant can result from slight elevation differences or insufficient slope. Using point clouds and elevation data makes it easier to identify ground surface inclinations and locations where water tends to accumulate, but if vertical accuracy is insufficient you may draw incorrect conclusions. When reviewing elevation data, check the height errors of validation points, the quality of ground surface extraction, the influence of vegetation, and the inclusion of structures, and verify that the accuracy meets the requirements for management purposes.

Cross-sectional checks are also effective. Settlement, heaving, changes in slopes, and scour around drainage channels that are difficult to discern from plan-view images become easier to grasp in cross-section. For example, cutting cross-sections along a maintenance road can reveal ruts and shoulder collapse. Cutting cross-sections perpendicular to a slope makes it easier to confirm tendencies toward collapse or bulging. However, cross-sectional data also depends on the quality of the point cloud, so you need to make careful judgments in areas where points are sparse or noise is high.

It is also important to view orthoimages and point clouds together. Orthoimages make visual changes easy to see, while point clouds allow you to confirm changes in height and shape. For example, areas where puddle marks or sediment outflow are visible in the images may not show clear elevation differences in the point cloud. Conversely, features that appear to be subsidence in the point cloud may be revealed by the images to be caused by weeds or shadows. Rather than judging based on a single result, cross-referencing multiple datasets increases the reliability of accuracy checks.

When comparing with past data, confirm whether the datasets can be compared under the same conditions. If flight altitude, imaging coverage, reference points, processing conditions, vegetation condition, or weather differ substantially between the previous and current surveys, what appears as differences may not represent actual change. For facilities in long-term operation, such as solar power plants, it is common to compare terrain and the conditions around equipment on a yearly basis. Therefore, when performing difference analysis, it is important to record differences in survey conditions and clearly define the ranges and caveats that are valid for comparison.

When checking deliverable data, we also look for unnatural gaps and missing areas. Under panels, in strongly shadowed areas, on water surfaces, in places with strong reflections, and in densely vegetated areas, point clouds and image processing may produce missing data. The presence of missing data is not necessarily a problem in itself, but if those gaps occur in locations that are important for management, additional verification is required. For example, if areas around drainage channels or the lower edges of slopes are missing, there will be insufficient information to judge poor drainage or sediment runoff. In such cases, consider supplementing with additional aerial photography, ground photographs, on-site surveying, or patrol records.

Drone surveying outputs provide materials that make it easy to grasp large areas in a short time, but they are not infallible. At solar power plants, panels, racking, shadows, reflections, vegetation, slopes, and drainage structures, among other factors, interact in complex ways. To verify surveying accuracy, it is necessary to compare the output data with on-site conditions and carefully examine any unnatural shifts or omissions, as well as the reasons for differences from past data. As point clouds and orthoimages are used more for management decision-making, the importance of this verification work increases.

Continuously Maintain Surveying Accuracy through On-site Verification and Record Management

Drone surveying is an effective method for managing solar power plants because it can cover wide areas from above. However, to make surveying accuracy useful in practice, it is necessary to combine on-site inspections with record management. If you try to make decisions based solely on drone surveys, you may overlook elements that are difficult to see in images or point clouds. In particular, drainage flow, muddy ground, scour beneath vegetation, small steps around structures, and minor cracks on slopes may only become apparent when confirmed on site.

In on-site inspections, we first extract locations of concern from the survey results. We target areas where the orthoimage shows color changes, where the point cloud shows unnatural changes in height, where there are large differences from past data, where drainage routes appear to be interrupted, and where the edges of management roads appear to be collapsed. By organizing the points to check in advance before going to the site, the efficiency of patrols increases and oversights are less likely.

On-site safety checks are also important at solar power plants. Spaces between panel rows, around mounting racks, slopes, drainage channels, muddy areas, and grassy ground are places that require caution when walking. Even when the purpose is to verify surveying accuracy, do not force entry into hazardous areas; instead, take photos and make notes within the range that can be safely inspected. If necessary, coordinate the permitted access areas in advance with managers and other relevant parties, and avoid touching equipment.

What is important in field verification is to correlate the drone survey results with how things actually appear on site. For example, confirm whether a spot that appears raised in the point cloud is actually overgrown vegetation, accumulated sediment, or the effect of mounting frames or materials. Areas that look dark on the orthophoto are also easier to judge on site—whether they are puddles, shadows, or wet ground. By recording these correspondences, you can make quicker determinations when checking results in the future.

In record management, we organize the survey date, flight conditions, reference points used, ground control points, checkpoints, imaging area, deliverable creation conditions, validation results, and on-site inspection results as a single workflow. In managing photovoltaic power plants, even if only the survey deliverables remain, if you do not know under what conditions the data were created, it becomes difficult to compare them later. Clarity of records is especially important when personnel change or when multiple stakeholders use the deliverables.

It is effective to record the results of survey accuracy checks not just as a simple pass/fail but also as usage notes. For example: the outer perimeter shows stable accuracy, but the ground surface beneath panels is difficult to see; the upper part of slopes can be confirmed, but the lower edges are affected by shadows; areas with dense grass require caution when treated as ground elevation. When such cautions are organized, users of the deliverables are less likely to make incorrect judgments.

Also, in periodic surveys, clearly identify changes since the previous survey. Record whether mowing was performed, whether earthwork repairs were carried out, drainage channel cleaning, maintenance or repair of access roads, surveying after heavy rain, effects of snow or freezing, whether any equipment construction was performed, and any other changes in site conditions. Without this information, when comparing survey results it becomes difficult to determine whether observed differences are due to actual topographic change, changes from maintenance work, or differences in imaging conditions.

To continuously maintain surveying accuracy at a solar power plant, it is important not to treat each survey as an isolated task. Establish a reference in the initial survey, check results in subsequent surveys using the same approach, and add points or revise imaging conditions as necessary. Because site conditions change over time, it is insufficient to fix a method once and for all. It is important to update the survey plan and the methods for verifying accuracy to match the plant’s terrain, management objectives, past failures, and maintenance history.

Record management also helps with internal sharing and explaining matters to stakeholders. When using drone survey results to explain the need for repairs, showing how accurate the data are, which locations were verified on site, and what changes have occurred compared with the past makes it easier to use them as a basis for decision-making. Conversely, if the basis for the results remains unclear, the survey data will remain merely reference material and be difficult to apply to practical decisions.

On-site verification and record management are the final steps that substantiate the accuracy of drone surveys. Even if reference points, flight plans, ground control points, and deliverable data have been checked, they will not fully realize their value as management documents unless linked to actual site conditions. In managing solar power plants, establishing a workflow of broadly capturing conditions with drone surveys, verifying them on-site, and using record management to inform subsequent surveys is the key to continuously improving survey accuracy.

Summary

To verify the surveying accuracy of a solar power plant using drone surveying, it is not enough to simply inspect the appearance of aerial images or point clouds. You need to comprehensively check whether control points and the coordinate system are properly established, whether the flight plan and shooting conditions are appropriate for the purpose, whether ground control points and check points are suitably placed, whether there are any unnatural misalignments in the point clouds or orthomosaic images, and whether on-site verification and record management have been carried out.

Solar power plants are sites where equipment is installed continuously across wide areas, and many factors—panel reflections, racking shadows, weeds, slopes, drainage channels, and access roads—can affect survey results. Therefore, it is essential to treat surveys not as mere extensions of general aerial photography but as surveying deliverables for operation and maintenance whose accuracy must be verified. In particular, when assessing drainage failures, ground subsidence, slope deformations, damage to access roads, or sediment runoff, reliability not only in horizontal (planimetric) position but also in the vertical (height) direction is important.

The strength of drone surveying is that it can efficiently record conditions over a wide area and preserve them in a form that is easy to compare with past data. On the other hand, the reliability of the results is determined by the accumulation of control points, imaging conditions, processing methods, validation, and on-site checks. If you systematize verification of surveying accuracy, it becomes easier to use the data as a basis for long-term power plant management decisions rather than as a one-off inspection.

When introducing drone surveying into the management of solar power plants, it is important to first clarify what you want to verify and to establish surveying conditions and methods for confirming accuracy that match that purpose. If you want to quickly grasp on-site changes and connect them to repair and maintenance planning, you need to design the workflow as an integrated series of tasks from surveying through result verification to record management.

To consistently verify the surveying accuracy of a solar power plant and apply it to daily management and maintenance decisions, it is important to continuously check not only the presentation of results but also the consistency of standards, validation, on-site verification, and records. When incorporating drone surveying into management operations, it is essential to organize accuracy conditions that match the purpose and establish an operational framework so that stakeholders can handle results on the same premises.