6 preparations for using drone surveying to create 3D models of solar power plants

3D models that provide a three-dimensional understanding of site conditions are useful for managing and designing solar power plants. When checking panel layout, graded surfaces, drainage routes, fences, access roads, slopes, and surrounding topography using only plans and photographs, one can misjudge distances, elevation differences, and areas with poor visibility.

Therefore, using drone surveying to record the entire plant from above and generate a 3D model makes it easier to communicate with stakeholders, compare before-and-after construction, and organize inspection records.

However, simply flying a drone and taking images does not instantly produce a usable 3D model. If you do not prepare a capture plan, ground control points, area coverage, image quality, overlap, data organization, and how to use the deliverables, you may later encounter problems such as “the necessary areas aren’t captured,” “it’s difficult to use for checking heights,” and “it’s hard to explain to stakeholders.” This article explains six preparations you should organize before using drone surveying to create 3D models of solar power plants, aimed at practitioners.

Table of Contents

• Clarify the purpose of creating the 3D model at the outset.

• Organize the target scope and the structures to be inspected

• Decide in advance how to handle reference points and coordinates.

• Tailor the shooting plan and flight conditions to the site

• Establish a system that can ensure image quality and overlap.

• Decide the verification items after data processing and the format of the deliverables.

• Precautions when using 3D models at solar power plants

• Summary

Clarify the purpose of creating the 3D model at the outset

When conducting drone surveys at solar power plants, the first thing to decide is the purpose for which the 3D model will be used. If you begin shooting with an unclear purpose, it may appear that flights went fine on site, but in later stages you can find yourself lacking the necessary accuracy, coverage, or viewing angles. Although the visual clarity of a 3D model is a major advantage, for practical use you need to first organize who will use it, what they will use it for, and which decisions it will inform.

For example, if the purpose is to survey a prospective site before new construction, the objective is to understand the terrain’s undulations, existing roads, surrounding trees, drainage direction, and the areas likely to require earthworks. In this case, it is necessary to capture the entire site and its connections to the surrounding topography more broadly. On the other hand, when used for operation and maintenance of an active power plant, inspection targets become more specific, such as around panel rows, beneath mounting racks, access roads, along fences, slopes, and drainage channels. If the purpose changes, the area to be photographed, flight altitude, and the deliverables that need to be checked also change.

The preparation required also varies depending on whether the 3D model will be used for design review, construction management, inspection records, or presentations to the client. For design review, the relative elevations of the terrain and the positional relationships with existing structures are important. For construction management, records are needed for parts related to before-and-after comparisons and the as-built condition. For inspections, it is important that abnormal areas can be reviewed later and that comparisons can be made with past conditions. For client presentations, a way of presenting that conveys the site conditions to non-specialists is required.

When setting objectives, it’s important not to rely solely on a 3D model. A 3D model produced by a drone survey is effective for understanding the overall positional relationships and topography of a site, but detailed dimension checks, inspections of equipment interiors, and electrical anomaly assessments need to be combined with other surveys or inspections. Defining the scope to be checked with the 3D model and the scope to be supplemented by on-site inspections, drawing reviews, and photographic records clarifies the role of the deliverables.

Also, the purpose of creating a 3D model needs to be shared among stakeholders. If the content that site staff, designers, contractors, management companies, and clients want to see differs, additional requests after capture are likely to arise. If additional capture is possible at the site, it may be possible to accommodate such requests, but re-capturing can be difficult due to weather, access restrictions, or operational conditions. Confirming the intended use in advance and setting priorities helps achieve drone surveying with less waste.

Organize the scope and the structures to be inspected

Next, organize the scope of the area to be 3D-modeled. Although a solar power plant may look like panels neatly aligned, it is actually made up of many elements such as site boundaries, graded surfaces, slopes, drainage facilities, maintenance roads, fences, access points (entrances and exits), electrical equipment, and the surrounding forests and farmland. If you shoot without deciding how far the survey should extend, you may lack necessary surrounding information or, conversely, end up collecting data for unnecessary areas, which increases the time required for processing and review.

When deciding the target area, it is important not to look only within the site but also to be aware of connections with areas outside the site. In a solar power plant, the flow of rainwater, inflow and outflow of sediment, connections with surrounding roads, boundaries with neighboring properties, and the location of perimeter fences are important management points. Even if the purpose of creating a 3D model is maintenance, including the outer areas a little more broadly can make it easier to trace causes later. In particular, slopes, drainage channels, and areas around fences are ranges that are often missed when only the power generation equipment itself is photographed.

Listing the structures you want to inspect before photographing them will make them easier to handle in practice. However, for the purposes of this text, the way of thinking is more important than the checklist.

Possible items include panel rows, mounting structures, related equipment such as power conditioners, maintenance roads, drainage channels, catch basins, detention ponds, perimeter fences, gates, slopes, retaining walls, site boundaries, existing trees, and surrounding roads. If you separate what you want to view in a 3D model for shape and positional relationships from what is sufficient to record in photographic detail, it becomes easier to create the shooting plan.

Also, because solar power plants have continuous rows of panels with the same shape, it can be easy to misidentify locations on the 3D model. Even if positions are known on-site, after data processing the screen often shows repetitive views, making it difficult to describe anomalous spots or areas that require inspection. Therefore, it is important to establish in advance criteria for locating places on the model, such as section names, aisle names, equipment numbers, cardinal directions, and positional relationships from entrances and exits.

If the area of interest is set too wide, the number of photos and the processing time will increase. Conversely, if it is set too narrow, there will be insufficient background information later for comparisons and explanations. When creating a 3D model of a solar power plant, it is important to include not only the power-generation equipment itself but also the areas relevant to management decisions. Before starting work, review site maps, existing drawings, past inspection records, and on-site photos, and clearly define the area that must be captured in this drone survey.

Decide in advance how to handle reference points and coordinates

When using 3D models in practical work, it is important not only to consider appearance but also how to handle position and elevation. 3D models created by drone surveys reconstruct three-dimensional shapes from overlapping images, so the usability of the deliverables depends on how you provide reference information. If you only need to grasp the three-dimensional appearance of a site, a simple alignment may suffice, but if you will use the model to overlay drawings, compare before and after construction, check heights, or assist with quantity calculations, you need to decide in advance how to handle reference points and coordinates.

Control points are important cues for aligning a 3D model to real-world coordinates and elevations. At solar power plants, because the site is extensive and similar rows of panels continue, a shortage of reference points can make overall model positioning and height verification unstable. If there are existing survey points on site, check whether they can be used. When establishing new calibration or validation points, consider placement, visibility, safety, workflow, and the impact on vehicle and pedestrian traffic.

What you should pay particular attention to is not to confuse the roles of control points and validation points. Control points are points used to align the model during creation, and validation points are points used to confirm the validity of the deliverables. If you use all points for alignment, it becomes difficult to assess how well the completed model matches independent points. When explaining deliverables in practice, separating the points used for model creation from those reserved for verification makes it easier to describe quality.

You also need to decide in advance how to handle the coordinate system. When overlaying existing drawings or design data, confirm which coordinate system they are managed in, what the elevation reference is, and whether the coordinates used on site match the coordinates on the drawings. If you create a 3D model while the coordinate system is ambiguous, it may look fine visually but can become misaligned when overlaid with existing data. In solar power plants, multiple drawings such as site development plans, equipment layout plans, drainage plans, and as-built drawings may be used, so it is important to decide which data to use as the reference.

Height also needs to be addressed based on the intended purpose. The required handling changes depending on whether it is sufficient to only observe terrain trends, whether it will be used to check slope faces or drainage gradients, or whether you want to see differences before and after construction. Vertical accuracy is affected by shooting conditions, control point placement, the shape of the target, and how the ground surface appears. Under panels or in areas with dense vegetation, the ground itself may not be captured adequately. Therefore, it is necessary to distinguish between heights that can be read from a 3D model and heights that should be confirmed by field surveying.

Preparation of control points and coordinates can sometimes be adjusted later, but if the necessary points are not visible in the images captured on site, it becomes difficult to address. Confirm the purpose of the survey, how the deliverables will be used, and the required positional accuracy, and establish a control point plan suited to the site; this will form the foundation for creating a reliable 3D model.

Adjust the shooting plan and flight conditions to match the site

In creating 3D models through drone surveying, the imaging plan greatly influences the quality of the deliverable. Solar power plants feature panels arranged regularly across a wide area, but they are also sites with many elements to watch for during capture—reflections, shadows, steps, slopes, fences, power lines, surrounding trees, and so on. If you shoot without fully considering flight altitude, shooting direction, and shooting intervals, you may end up with insufficient image overlap, obscured details, or portions of the model becoming distorted.

First, flight altitude is determined by the balance between the level of detail you want to see and the size of the area to be covered. Raising the altitude lets you efficiently capture a larger area, but ground details become harder to see. Lowering the altitude makes it easier to capture fine details, but increases the number of images, as well as flight time and data-processing workload. Separating flights for capturing the overall terrain and layout of the power plant from flights for checking details such as areas around fences, slopes, and drainage facilities makes it easier to produce deliverables that meet the objectives.

The direction of capture is also important. Straight-down images produce a view close to a plan, making it easier to grasp the overall layout and the continuity of the ground surface. Conversely, combining oblique images can better convey the three-dimensionality of racking, slopes, retaining walls, fences, and areas around equipment. In 3D models of solar power plants, panel surfaces tend to reflect light and can produce repetitive patterns, so it is important to vary shooting angles as needed. However, if you increase oblique shots, you need to carry out more thorough checks for nearby obstacles and the safety of flight paths.

In flight conditions, consider the effects of weather, wind, solar altitude, shadows, and reflections. Clear skies make scenes appear brighter, but reflections on panel surfaces and shadows can become prominent. Cloudy skies can soften shadows, but attention is needed to image contrast and how the ground surface appears. On windy days, the aircraft’s attitude can become unstable, leading to image blur and variation in capture positions. In rain, fog, or conditions with strong backlighting, you should judge whether to continue shooting based on the potential impact on the deliverables rather than forcing the operation.

At solar power plants, safety checks along the flight route are also indispensable. Confirm the perimeter fence, utility poles, overhead lines, surveillance equipment, surrounding roads, adjacent land, and the flow of people and vehicles, and secure takeoff and landing areas safely. Inside the plant, internal access roads may be limited, and inspections or maintenance work can coincide with flights. It is important to coordinate work times with on-site personnel and ensure that no one enters the flight area.

A shooting plan is not sufficient if it only establishes the flight route. Deciding in advance what area to capture at what resolution, which parts to supplement with oblique shots, and which locations to prioritize if reshoots become necessary will make it less likely you’ll be uncertain on site. Because solar power plants are often large sites, discovering later that you missed captures can lead to significant rework. In the planning stage it is important to consider purpose, coverage, safety, and quality together.

Create a system to ensure image quality and overlap

The quality of a 3D model is largely determined by the quality of the source images. If images are blurred, too dark, too bright, do not have sufficient overlap of the same areas, or have the subject partially cut off, there are limits to what can be corrected in post-processing. In drone surveys of solar power plants, reflections from panels, regular patterns, the texture of grass or gravel, and shadows on slopes can all affect results, so it is important to have a system in place to ensure usable images at the time of capture.

Image overlap is a particularly important factor in creating 3D models. When adjacent images share sufficient common areas, the processing can recognize the same features and more easily reconstruct the three-dimensional shape. If overlap is insufficient, the model may develop holes, its positioning may become unstable, or some shapes may become distorted. At solar power plants, long runs of similar panel rows can make matching between images difficult. Therefore, it is important—more so than at typical sites—to pay attention to capture overlap and the continuity of the route.

To ensure image quality, pre-shoot checks are essential. Check for lens dirt, incorrect settings, recording capacity, battery status, and the recording status of time and location information. When you review images after shooting, if they are generally blurred or the exposure is extreme, retakes may be necessary. By establishing procedures for simple on-site checks, you can detect missed shots or insufficient quality earlier.

Solar panels are prone to reflections, which can cause blown-out areas in images. Blown-out regions make it difficult to read surface features and can be disadvantageous for modeling and condition assessment. It is not appropriate to perform detailed degradation assessment of the panels themselves using only a 3D model, but measures to avoid excessive reflection are useful for understanding the positional relationships of panel rows and surrounding structures. Take into account the time of day, shooting angle, and weather to obtain as stable images as possible.

Also, on sites with a lot of grass and shrubs, the ground surface may not be captured in the images. Because the surface of the vegetation is reconstructed as geometry in the 3D model, it may not match the actual ground surface. If you use it to check drainage direction or subsidence, you need to be aware of this. When comparing models before and after mowing, differences in vegetation condition rather than actual terrain changes may appear. As needed, it is important to combine site preparation before imaging and on-the-ground verification.

As a quality assurance measure, it is reassuring to assign roles not only to the pilot but also to a site inspector, a safety spotter, and a data verifier. Even on small sites, it is a heavy burden for one person to check image quality, surrounding safety, and target coverage while flying. Assign roles according to site conditions and establish a workflow that allows you to immediately check image count, target coverage, blurriness, brightness, and the capture of reference points after shooting; this will reduce failures in 3D model creation.

Determine the verification items and deliverable formats after data processing

After the shoot is finished, the images are processed to create deliverables such as 3D models, point clouds, and orthophotos. However, simply completing the processing does not mean they are ready for practical use. To make use of a solar power plant’s 3D model, you need to decide in advance what will be checked after completion, in what format it will be handed over to stakeholders, and how much explanation will accompany it.

The first thing to check is whether the target area has been reproduced without omission. Verify that the locations you decided beforehand as important—site perimeter, along fences, entrances/exits, maintenance roads, slopes, drainage facilities, and the ends of panel rows—are included in the model. Even if the model looks clean, it is insufficient as a deliverable if the essential inspection points are missing. In particular, perimeter areas and places with elevation differences tend to fall at the edges of the imagery and can have sparse data.

Next, carry out checks using control points and check points. When using positions or elevations for practical decision-making, it is important to verify how well the points in the model match the on-site check points and to be able to explain the scope of use of the deliverables. The important thing here is not to treat a 3D model as an all-purpose surveying result. Confirm that the quality is sufficient for the intended use, and if detailed dimensions or elevations need to be handled precisely, combine it with other survey results or on-site verification.

The form of deliverables also varies depending on the purpose. For explaining the overall situation, overhead orthophotos and simple 3D visualizations can be easier to understand. For checking terrain, slopes, and developed surfaces, point clouds and data that allow confirmation of elevations are effective. For presentations to stakeholders, images from representative viewpoints, annotated materials, and comparison screens are useful. If clients or managers are not accustomed to specialized operations, it is also important to organize the materials into easy-to-use formats.

Also, because 3D model deliverables tend to have large file sizes, you should also prepare a method for sharing them. Decide who will store them, which format will be treated as the official deliverable, whether to save the original images before editing, or whether to keep only the processed data. If there is a possibility of reprocessing or comparison later, it is convenient to organize and save the original images, control point information, shooting conditions, processing conditions, and the date the deliverable was created.

When reviewing deliverables, it is important not to be overly swayed by their visual impact. Because 3D models can present a site intuitively, they leave a strong impression on stakeholders. However, they can also contain shape distortions caused by shadows, reflections, vegetation, or insufficient imaging. When explaining the deliverables, you should communicate separately what can be confirmed and what cannot. For example, while the general terrain undulations and the positional relationships of structures are easy to grasp, you may not be able to determine precise ground elevations where the surface is covered by grass or the condition of interiors of equipment.

By deciding in advance the checklist for post-processing and the format of deliverables, you can convert drone surveying results into information that is easy to use on site. It is important to regard the process not as ending with the capture, but as a series of preparations that includes checking, organizing, sharing, and linking the results to subsequent decisions.

Considerations When Using 3D Models at Solar Power Plants

The 3D model of a solar power plant is a useful deliverable that makes it easier to understand the entire site, but if used incorrectly it can lead to misjudgments. In particular, because it appears three-dimensional and easy to interpret, caution is needed since it can give the impression that everything is represented accurately. 3D models obtained from drone surveys are created based on the captured images, site conditions, and processing methods, and there are limits to what can be verified.

First, areas that are shaded under panels or mounting racks may not be adequately recorded by aerial photography alone. At solar power plants, wiring, support members, the ground surface, and drainage flow beneath panels can be involved, yet aerial imagery may not reveal these areas. It is risky to assume that parts not visible on a 3D model are in good condition. As needed, it is important to combine with ground photos and on-site inspections.

Next, there are the effects of vegetation and temporary obstacles. Grass, scrub, materials, vehicles, temporary structures, and people moving through the area can be captured during image acquisition and may therefore be represented in the model. In grassland in particular, the grass surface—not the ground—is rendered in 3D, so care is required when checking ground shape. When making pre- and post-construction or seasonal comparisons, recording the time of capture, mowing status, weather, and the way shadows fall makes it easier to interpret the meaning of any differences.

Also, consistency of conditions is important when comparing 3D models. If flight altitude, shooting angle, ground control points, processing parameters, or the target area differ greatly between the previous and current acquisitions, differences in appearance can easily look like actual changes. In the operation and maintenance of solar power plants, there are cases where you want to periodically compare the same location. In such cases, it is desirable to capture imagery under as similar conditions as possible and to establish procedures to produce outputs according to the same standards.

Furthermore, when explaining to stakeholders, consideration must be given to how the 3D model is presented. Technical specialists can check point clouds and coordinate information, but clients and management staff may not be familiar with navigating the interface. For explanatory purposes, combining an overall view, problem areas, zoomed-in views, spatial relationships, and on-site photographs makes the content easier to understand. Rather than simply handing over the 3D model, it is important to clarify what to look at and which points are the decision-making criteria.

In solar power plants, coordination with the operation of power generation equipment and maintenance work is also necessary. Drone surveying is a means of streamlining on-site tasks, but the flights themselves can affect site management and safety management. Access control, flight zones, consideration for nearby residents and adjacent properties, notifying workers, and emergency response should be confirmed in advance. Operations must be carried out with safety as the top priority.

When using 3D models, also consider the storage of deliverables and update management. A model once created becomes a record of the site conditions at that time. If, when you review it later, you cannot tell when, over what area, or under what conditions it was captured, it becomes difficult to use as comparative material. Organize the file name, capture date, target area, work purpose, reference point information, and verification results, and make sure they are linked to the next survey.

Summary

When using drone surveying to create 3D models of solar power plants, the important thing is not to focus only on flight and photography but to prepare the entire workflow from goal setting to the utilization of the deliverables. 3D models clearly show the site’s three-dimensional conditions and help align stakeholders’ understanding, record inspections, and compare before and after construction. On the other hand, if the coverage area, control points, image quality, data processing, and sharing methods are insufficient, the deliverables may look tidy yet be difficult to use for practical decision-making.

First, clarify what the 3D model is being created for. The information required differs for design review, construction management, maintenance, and client briefings. Next, organize the target scope and the structures you want to check, and decide whether to include not only the area around the panels but also the perimeter fence, slopes, drainage facilities, access/maintenance roads, and surrounding topography. Furthermore, decide how to handle control points and coordinates, and clarify the relationship with existing drawings and site surveys.

In the imaging plan, flight altitude, shooting direction, image overlap, weather, reflections, shadows, and safety management are considered according to site conditions. To ensure image quality, procedures for pre- and post-acquisition checks are also necessary. After processing, verify the level of reproduction of the target area, consistency with control points, and the usability of the deliverables, and organize them into a format that stakeholders can easily use. Finally, it is important to explain separately what can be confirmed with the 3D model and what should be supplemented by on-site inspection or additional surveys.

Drone surveying of solar power plants is not merely a recording task; it is a means to organize site conditions clearly and guide subsequent decisions. If you prepare with 3D model creation in mind, it becomes easier to share information during inspections, consultations, design verification, and maintenance management. If you want to understand site conditions three-dimensionally and put survey results to practical use, it is important not to rely solely on specific equipment or service names, but to establish a drone surveying approach that fits your company’s management objectives, required accuracy, and operational structure.