Can As-Built Drawings Be Created from Point Clouds? 5 Foolproof Steps to Create Them

On projects that produce current-condition documentation, practitioners have traditionally combined plane-table surveying, total station observations, photographic records, and manual drafting to carefully record shapes and positional relationships. These methods are still effective, but when the subject is complex, when a wide area must be recorded in a short time, or when there is a need to review details later, relying solely on conventional techniques can become burdensome. This has brought attention to the use of point clouds.

A point cloud is data that captures the surface of an object as a large number of three-dimensional coordinates, and its major characteristic is the ability to record the on-site shape at high density. For those responsible for creating as-built preservation drawings, the ability to later check details that are easy to overlook in the field, to examine cross-sections or elevations as needed, and to retain records even in locations that are difficult to revisit are significant advantages. On the other hand, simply collecting a point cloud does not produce a drawing. If several key points—such as the concept of accuracy, alignment with reference coordinates, interpretation of the features to be mapped, and decisions about vectorization and representation—are not addressed, the data can actually become difficult to use.

This article first clarifies whether it is truly possible to create as-built drawings using point clouds, then explains a 5-step creation procedure that reduces the likelihood of failure in practical work. It also carefully summarizes, from the perspective of practitioners, common pitfalls when generating drawings from point clouds and approaches to streamlining coordinate verification on site.

Table of Contents

• Can as-built preservation drawings be created from point clouds?

• Basics to understand before creating as-built preservation drawings from point clouds

• Fail-proof creation procedure 1 Define the purpose and deliverables

• Fail-Proof Creation Procedure 2: Designing Acquisition Conditions

• Foolproof creation procedure 3: Organize point clouds and align the reference

• Foolproof creation procedure 4 Read the information necessary for diagramming

• Fail-proof creation procedure 5: Organize as an as-is preservation drawing

• Common mistakes when creating as-built preservation drawings from point clouds

• Considerations for Stabilizing the Creation of As-Built Preservation Drawings Using Point Clouds

• Summary

Can as-built drawings be created from point clouds?

In short, existing-condition preservation drawings can be produced from point clouds. However, to be more precise, point clouds are an extremely powerful foundational dataset for creating existing-condition preservation drawings, and the point cloud itself does not directly become the finished drawing. Only by extracting the necessary shapes from the point cloud acquired on site, converting them into drawings, and formatting the presentation to suit the recording purpose does it become usable as an existing-condition preservation drawing.

The important point here is not to misunderstand the role of current-condition record drawings. A current-condition record drawing is not a mere visual reproduction; it is an organized record that presents the object's position, outline, elevation differences, extent of damage, relationships with the surroundings, and so on so that a third party can understand it later. For that reason, point clouds with a large amount of information are certainly powerful, but because they contain so much information they can be difficult to read as-is, and the meaning of the necessary lines and surfaces may not be conveyed. Being able to bring back a high-density representation of site conditions and having that information communicated as a record drawing are similar but distinct matters.

Point clouds remain effective because they can consistently preserve subtle undulations and complex shapes over a wide area that traditional spot-sampling surveys tend to miss. In particular, features such as the undulations of archaeological remains, the surface geometry of stone masonry, deformation of existing structures, traces of collapse or damage, and terrain differences before and after construction are difficult to represent with methods that sample only a few survey points. With point clouds, locations that went unnoticed in the field are easier to check during post-processing, and not only plan views but also cross-sections and elevation drawings can be extracted as supplementary materials.

On the other hand, using point clouds does not automatically yield high-quality preservation record drawings. If accuracy is lacking, the relationship to control points is unclear, acquisition density is insufficient, or many extraneous objects are included, the drawing process can actually become more difficult. In short, the practical answer is that preservation record drawings can be made from point clouds, but they will not succeed unless the workflow—from point cloud acquisition through drawing—is designed to match the documentation objectives.

Fundamentals to Grasp Before Creating As-Built Preservation Drawings from Point Clouds

When creating preservation drawings from point clouds, you first need to make clear "what the drawing is meant to preserve." Even for preservation drawings, the information to prioritize varies depending on the subject. Whether you want to carefully record shapes and damage conditions as with archaeological sites and cultural properties, understand the layout and dimensional relationships of existing structures, or record the undulations of slopes and the ground surface, the required point cloud density, acquisition directions, method of establishing references, and the way drawings are represented will vary greatly.

Next, what you should understand is that a point cloud contains the site as it was at the time of capture. This is both an advantage and a disadvantage. For example, depending on the conditions, workers, temporary structures, vegetation, vehicles, protective coverings, puddles, and highly reflective surfaces that have a significant effect may also be recorded. While this is valuable for faithfully preserving site conditions, unnecessary elements can obstruct interpretation during the drawing/diagramming stage. Therefore, it is important to understand that point clouds are not an all-purpose correct dataset, but data that should be handled on the premise of selection and organization.

Furthermore, what is required for an as-found preservation drawing is the judgment to translate three-dimensional information into a two-dimensional drawing. For example, decisions such as which parts of the surface relief to adopt as contour lines, where to delineate the boundaries of damage or loss, and to what degree to depict surface irregularities cannot be decided by point-cloud density alone. Taking into account the scale of the final deliverable, its intended use, and the ease with which viewers can understand it, you must design how much to represent.

If you keep these basics in mind, you'll reduce waste starting at the point cloud acquisition stage. Conversely, if you collect point clouds "just in case" while the purpose remains unclear, you are likely to end up with data that is large in size but difficult to visualize or convert into drawings. When creating as-built drawings, it's more important to determine acquisition conditions by working backwards from the deliverables than to focus on the measurement itself.

Foolproof Creation Procedure 1: Define Objectives and Deliverables

The first step is to clarify what the current-condition record drawings are for. If this is vague, the subsequent methods for acquisition and processing will become inconsistent. For example, the required density of records will vary depending on whether the focus is on capturing plan positions, ensuring reproducibility of cross-sections, or documenting the damage condition of elevations in detail. Likewise, the items that need to be prepared will differ depending on whether the deliverable is a paper drawing or whether three-dimensional data will also be used in later stages.

In the practical work of creating as-built preservation drawings, it is extremely important to define up front the scale and level of detail required for the finished drawings. If the purpose is a broad overview, it may be sufficient that the overall spatial relationships and contours are consistent. On the other hand, if you want to record missing details or the condition of joints, the required resolution can vary greatly even for the same point cloud. A point cloud may seem to allow you to extract anything later, but in reality information does not increase beyond the limits of the acquisition density and viewpoint conditions. Therefore, you need to decide in advance how much detail you will depict.

Also, at this stage we decide what reference to use for the drawings. Whether to manage them in site coordinates or whether arbitrary coordinates are acceptable, which elevation datum to adopt, and whether consistency with existing drawings and other deliverables is necessary are the points to consider. Even if the record drawings alone appear fine, if positions do not align when overlaid with other survey results or design drawings, it will lead to major rework downstream. In practical point cloud work, the consistency of the coordinate system and references affects later usability more than how tidy the visual appearance is.

Additionally, organize the supplementary records that should be used on site. For example, differences in material, color boundaries, damage that can only be identified by visual inspection, and the condition of buried features or backfill, which are difficult to judge from point clouds alone, need to be supplemented with photographs and notes. If you try to rely entirely on point clouds, the basis for decisions will be insufficient at the drafting stage, and delineation tends to become ambiguous. Clarifying the purpose and the deliverables is important not only for determining the measurement method but also for deciding how to collect supplementary information.

Fail-Proof Creation Procedure 2 Designing Acquisition Conditions

The next step is to design the conditions for point cloud acquisition according to the purpose. By "design" here we mean not only which equipment category to use, but also from where, at what density, over what area, and according to what criteria the data will be captured. For as-built preservation drawings, it is more important that the surfaces and boundaries you want to depict can be read than merely that points have been captured. Therefore, simply increasing point density is not always the right approach.

For example, the appropriate acquisition direction differs between cases where you want to cover a wide area dominated by planar surfaces and cases where you want to read the surface shapes of walls or stone masonry in detail. There are situations where acquisition from overhead or from above is advantageous, and others where a side perspective is necessary. For objects with many blind spots, measurements from only one direction do not provide sufficient information. In practical work creating as-built preservation drawings, it is essential at the point cloud acquisition stage to first consider "which surfaces will become unseen."

Also, the effects of site conditions cannot be overlooked. Confined spaces, strong sunlight, wet surfaces, highly reflective materials, vegetation cover, and the movement of people and vehicles can greatly affect point cloud quality. Depending on on-site conditions, it may be better to acquire the data in multiple sessions. In particular, if contours that should be preserved in as-built drawings are hidden by grass or temporary materials, post-processing may not be able to fully recover them. Before entering the site, considering whether unwanted items can be moved and whether work hours can be adjusted will reduce the effort required after acquisition.

Even more important is tying the data to control points and known points. Even if the point cloud is clean, if the references are ambiguous the reliability of the deliverables is reduced. When integrating multiple measurement datasets or when there is a possibility of re-surveying later, it is necessary to leave on-site references that can be verified. If this is omitted, each point cloud may be internally consistent yet difficult to compare with others or to align with drawings. As-built preservation drawings have meaning in "preserving the current condition in a form that can withstand future comparisons," so reproducible reference management is indispensable.

Foolproof creation procedure 3: Organize point clouds and align the reference

The acquired point cloud should not be taken straight into drafting; it must first be cleaned up and aligned to a reference. If this step is lax, misalignments and reading errors on the drawings are more likely to surface later. Even point clouds that seemed dense on site can, if there is a slight tilt or positional offset during data integration, cause contour lines to appear doubled or sections to become thick and blurred, degrading the clarity of the as‑found documentation.

The first step in cleanup is removing unnecessary points. Remove people, vehicles, temporary structures, vegetation swayed by wind, stray points, noise, and so on as needed to make the target’s outline easier to read. However, be careful not to remove too much. In current-condition preservation drawings, the state of the site itself may be the subject of documentation. For example, temporary protective materials that covered damaged areas or the spread of collapsed soil may need to be recorded. To avoid confusing unwanted objects with subjects to be recorded, editing should be done by returning to the purpose defined before acquisition.

Next is aligning multiple datasets and assigning coordinates. For extensive sites or complex subjects, it is common to integrate data acquired multiple times. At this stage, it is not enough that the pieces visually join naturally; you must confirm they are consistent with the reference. Even if they match locally, the whole can be slightly twisted, and that may only become noticeable when producing plans or sections. Therefore, before diagramming you should compare with known points or control points, check for offsets on sections, and verify the key dimensions.

Also, it is necessary to have a perspective for organizing data to a density suitable for visualization. Point clouds are often thought to be better the denser they are, but extremely heavy datasets can impair usability and actually make it harder to interpret important contours. If you prepare the data so that the display can be switched according to purpose—such as for overall inspection, detailed inspection, or cross-section extraction—the efficiency of visualization improves. Since as-built drawings are ultimately artifacts that people will read and understand, it is important to organize the data-processing stage with an awareness of "readability."

Foolproof Creation Step 4: Read the Information Necessary for Diagramming

Once the point cloud cleanup is finished, the next step is to enter the process of extracting the information necessary for drafting. This is the core that separates mere 3D data processing from the creation of as-built preservation drawings. A point cloud contains a vast number of points, but what is needed for drawings are meaningful outlines, changes in surfaces, boundaries, elevation differences, areas of damage, and so on. In other words, rather than the number of points, it is important to be able to decide "what to take as lines."

The first thing to be aware of is that the main elements of as-built preservation drawings are the relationships between lines and surfaces. When looking at point clouds, it's easy to be drawn to the details, but what a drawing needs to convey is the skeleton that communicates the composition of the subject. For example, prioritize capturing the information a drawing reader should grasp first — such as the top edge, ends, bend points, slope shoulder and slope toe, transitions in wall faces, outlines of defects, material boundaries, and locations where elevation changes. Detailed representation comes afterward. If you reverse this order, you often end up with drawings that contain a lot of information but make it hard to see the essential structure.

Also, it is important not to judge based only on plan views. In as-built drawings, changes that are not visible in plan may only become apparent in cross-sections or elevations. For example, an edge that looks straight on the plan may be wavy in a cross-section, and an elevation can reveal the depth of damage or the tilt of a surface. The strength of point clouds is that you can change viewpoints to check details. Therefore, the person responsible for drafting should determine the position of lines by moving between plan, cross-section, elevation, and overhead views as necessary.

Furthermore, consistency is essential in how lines are traced. If the reference shifts by location for boundaries of the same type, the overall reliability of the drawing is reduced. In particular, for objects with many irregularities or whose boundaries have become ambiguous due to aging, it is important to decide in advance which surface will serve as the representative value when converting to a line. Drawing lines based on local impressions breaks continuity and causes practitioners to become confused when using the drawings later.

Point clouds are extremely useful as material for rechecking when you cannot return to the site, but they do not automatically substitute for the judgment involved in drafting. That is why it is important for someone who can read point clouds to work with both an understanding of the site and an understanding of the deliverables. As-built preservation drawings are both record drawings and drawings meant to communicate to the reader. Therefore, linework that is organized so its meaning can be discerned—not merely precise—is required.

Foolproof Creation Step 5: Prepare as an As-Built Drawing

The final step is the process of organizing the information you have captured into an as-built drawing. Here you improve the drawing’s readability, usability, and explanatory quality. Even if you have been able to trace lines in the correct positions from the point cloud, they are often hard to read as-is. By refining line types, annotations, section locations, necessary legend-like explanations, and removing extraneous information, you increase the completeness of the deliverable.

In condition-record drawings, drawing everything in is not always the right approach. Instead, it is important to vary the emphasis so that the information that needs to be conveyed does not become buried. Use lines that show the main contours, boundaries, and level differences as the backbone, and narrow down detailed information according to the purpose. For example, representing every small surface abrasion with lines can obscure the overall composition. Conversely, elements important for preservation records, such as losses or cracks, should be prioritized and shown clearly. The judgment of what to omit and what to retain determines the quality of the drawing.

Also, be mindful that the drawings are derived from point clouds, and it is important to word them in ways that are supported by evidence. While cross-checking with on-site reference photos and notes, avoid making unwarranted assumptions in areas where judgments may differ. Overfilling areas with sparse points for appearance’s sake will undermine the reliability of record drawings based on actual measurements. Not leaving unclear parts ambiguous is also a condition of a good existing-condition preservation drawing.

Furthermore, organizing the deliverables with their post-submission uses in mind will be useful in practice. For example, if you provide not only plan drawings but also indicate the locations of required sections, it will be easier to carry out additional reviews later. It is also important to organize the reference information and the logic behind the coordinates so that another person can understand the criteria used when the drawings are reused. As-built drawings are often not simply submitted and finished; they frequently become baseline materials for future comparisons and repair evaluations, so it is worthwhile to make them easy to reuse.

Common Mistakes When Creating As-Built Drawings from Point Clouds

One common mistake when creating as-built drawings using point clouds is insufficient acquisition coverage, resulting in missing necessary surfaces. Even if it appears on site that everything was captured, when reviewed later there can be many blind spots and the critical edges or break points may be unclear. Pay particular attention to areas along walls, behind steps, confined spaces, and parts hidden by vegetation. This is caused not only by the performance of the equipment but also by inadequate pre-acquisition planning.

Another common problem is proceeding while the coordinate reference is left ambiguous. Even if a point cloud is internally consistent, if its relationship to external references is unclear it cannot be compared with other drawings or with previous years’ deliverables. Because as-built drawings are often produced for preservation and comparison, this mistake can have major repercussions later. Do not postpone verification of control points; they should be considered together from the outset.

There have also been failures caused by overreliance on point cloud data and by taking on-site verification lightly. Point clouds are extremely useful, but differences in material, color boundaries, distinguishing soft vegetation from hard structures, and assigning meaning to surfaces can be difficult to judge without on-site observation or photographs. If you try to complete everything using only point clouds, you can become unsure during the drafting stage about "what exactly is included."

Furthermore, a lack of organization in drawings is often seen. A common example is cramming in data derived from point clouds as-is, resulting in diagrams that are difficult to read because there are too many lines. I understand the desire to improve recording accuracy, but as-built drawings must, above all, communicate to the reader. Adding more information is not the same as making the drawing easier to understand. You need to create layers of information according to purpose and design the drawing to bring the primary information to the forefront.

Approach to Stabilizing the Creation of As-built Preservation Drawings Using Point Clouds

To consistently produce as-built preservation drawings from point clouds, it is important not to divide surveying/acquisition, processing, and drafting into separate stages. Even if the person who collects data on site, the person who processes the data, and the person who creates the drawings are different, sharing the purpose of the deliverable can reduce omissions of necessary information. Conversely, if the acquisition team focuses only on capturing points and the drafting team proceeds on the assumption they will fix things later, problems will inevitably arise.

In that sense, the most important idea is to "work backwards from the drawing." First envision the contours, dimensional relationships, sectional representations, and comparison targets required in the finished drawing, and then determine the point-cloud quality and reference control needed for that. If this flow is in place, point clouds become an extremely powerful foundational resource. Although point clouds can contain a large amount of information, if handled without a purpose they tend to become unmanageable data. That is precisely why defining the purpose and the deliverables is the top-level quality control.

Also, having a system that can carry out quick and accurate on-site position checks directly affects overall stability. When checking reference points, adding supplementary points, and reacquiring the positional information used as references during mapping can be done smoothly, the connection between point clouds and drawings becomes stronger. In particular, when dealing with large outdoor sites, if on-site coordinate checks take time, both data acquisition conditions and mapping accuracy tend to become unstable.

In such situations, it is advantageous to have means to streamline checking site coordinates and identifying control points. For example, a GNSS high-precision positioning device such as an LRTK that can be attached to an iPhone makes it easier to perform centimeter-level position checks (cm level accuracy; half-inch accuracy) on site while aligning the reference for point cloud acquisition and verifying auxiliary points. In creating as-built drawings, even if the final deliverable is a drawing, the way site coordinates are handled determines quality. By combining LRTK not only for point cloud acquisition itself but as a means to streamline control point surveying and site coordinate checks, it becomes easier to iterate between the field and mapping, making it easier to achieve both overall work accuracy and speed.

Summary

Existing-condition drawings can be created sufficiently with point clouds. In fact, for sites with complex shapes, large areas, or where detailed checks are needed later, point clouds serve as a powerful foundational resource that strongly supports conventional methods. However, acquiring point clouds does not automatically produce good drawings. Only when you clarify what the drawing is intended to record, design the acquisition conditions, align reference points, interpret meaningful contours, and organize them into drawings as a whole does an existing-condition drawing become usable in practice.

Most importantly, do not be overwhelmed by the amount of point cloud data; identify the information required for the deliverables. Current-condition record drawings must not only be precise but also readable and understandable later, and usable for comparison and review. To achieve this, measurement accuracy, coordinate alignment, decisions about depiction, and drawing representation all need to be considered together.

If you want to make checking reference points and understanding on-site coordinates more efficient, using LRTK can be effective. As an iPhone-mounted high-precision GNSS positioning device, it makes it easy to perform centimeter-level position checks on site (cm level accuracy, half-inch accuracy) and helps streamline verification of control points and coordinate management before and after point cloud acquisition. To stably produce as-built preservation drawings from point clouds, it is important not only to rely on drafting software but also to organize how coordinate operations are handled on site. On projects that want to leverage the recording accuracy of point clouds while reducing the effort of on-site verification, it is worth considering an operational workflow that combines LRTK.