8 Tips for Generating Plan Views and Longitudinal Profiles from Buried Pipe Point Clouds｜Including Cross-Section Extraction

Introduction

Accurately recording the locations of buried water and sewer pipelines and gas pipes and drafting them as plan views and longitudinal sections is an important task for post-construction maintenance and for preventing interference with other works. Traditionally, surveying staff measured the positions and depths of buried pipes point by point and manually prepared the drawings. However, this method limits the number of points that can be collected and can fail to fully capture changes in pipe gradient or clearances from surrounding features. Recently, attention has turned to the use of point cloud data. Using point clouds acquired by laser scanners or LiDAR allows the shapes and alignments of buried pipes to be recorded at high density and high accuracy, enabling detailed analysis from cross-sectional shapes to longitudinal profiles. However, creating plan and longitudinal drawings from raw point clouds requires several specialized procedures and techniques. Here, we introduce eight tips for using point cloud data of buried pipes in practice to produce drawings. If pipelines inside a trench (open excavation) are scanned, their accurate plan positions and depths can be determined from those point clouds even after backfilling. Below, while touching on diverse measurement methods such as terrestrial laser scanners, UAV-mounted LiDAR, smartphone LiDAR, and RTK-GNSS receivers, we explain concrete points for generating plan views and longitudinal sections of buried pipes through point cloud processing.

Tip 1: Choosing the Right Measurement Method and Timing

First, plan how to acquire the point cloud data. Depending on the condition of the buried pipes and the site environment, it is important to choose the optimal measurement method. Typical methods include terrestrial 3D laser scanners (TLS), drone-mounted LiDAR, smartphone-built-in LiDAR, and RTK-GNSS surveying. Each has its own characteristics; for example, stationary laser scanners can acquire high-density point clouds with millimeter-level accuracy, but the equipment is large and setting up and measuring at multiple locations is time-consuming. UAV (drone)-mounted LiDAR can scan wide areas from the air in a short time, making it effective for understanding the terrain of long pipelines and areas that are difficult to access. However, drones have difficulty capturing fine details inside deep trenches, and flight restrictions and visibility limitations due to trees and buildings must be considered. Meanwhile, the increasingly common smartphone LiDAR function has the advantage of being easy to carry and can capture data even in narrow sites. Smartphone LiDAR has an effective range of about 5 m (16.4 ft), so it is not suitable for surveying wide areas, but it is well suited for finely scanning local areas such as trenches for buried pipe work and around pits. By scanning from multiple positions as needed, smartphone LiDAR can cover a certain extent. Furthermore, surveying with an RTK-GNSS receiver can obtain point coordinates with an accuracy of several centimeters (cm level accuracy (half-inch accuracy)), so it is effective for surveying control points and confirming key positions. For example, measuring the start and end points of long pipelines, curve points, and points that serve as vertical reference with GNSS makes later drafting and accuracy verification easier.

Timing of measurement is also an important factor. For buried pipes, it is ideal to perform point cloud measurement during open-cut work while the piping is exposed as much as possible. Because once the pipe is backfilled it cannot be measured directly by laser, scan before and after burial to keep a record. For example, if you 3D-scan inside the trench after the piping has been installed, you can comprehensively capture the pipe’s alignment (planar route) and the elevations of the top and bottom. Scanning only the ground surface later will not show the pipe itself in the point cloud, so the key is to perform measurements during construction without missing the timing. As described above, by selecting the appropriate equipment according to site conditions and acquiring point clouds at the timing that yields the most information, you can prepare high-quality source data.

Tip 2: Ensuring a High-Accuracy Coordinate Reference

When generating drawings from point clouds, the coordinate accuracy of the point cloud data and alignment to the reference are critically important. No matter how detailed the point cloud is, if it does not match the site’s survey coordinate system, it cannot be accurately reflected in plan views or longitudinal profiles. Therefore, it is essential to implement measures to acquire absolute coordinates with high precision from the measurement stage.

One method is to combine measurements with RTK GNSS. By using RTK-GNSS, you can append latitude, longitude, and elevation information to the acquired point cloud in real time, yielding a point cloud directly tied to the global geodetic reference frame. For example, even when scanning with smartphone LiDAR, attaching a small RTK-GNSS receiver to the phone and synchronizing both sets of data makes it possible to assign geodetic coordinates to each point. If the point cloud is georeferenced (spatial coordinate alignment) at the acquisition stage in this way, the effort of transforming it to match reference points in post-processing can be greatly reduced. The same applies to drone LiDAR: by equipping the aircraft with high-precision GNSS and an inertial measurement unit and calibrating beforehand with known points, the output point cloud can directly reflect map coordinates.

If GNSS signals are difficult to receive in a location (such as in the shadow of buildings in urban areas, in mountainous regions, or indoors), the conventional method of installing known control points on site and using them to align point clouds is effective. For terrestrial laser scanners, multiple spheres or target plates called targets (control points) are placed, and their coordinates are measured with a total station or GNSS. By matching the target positions in the resulting point cloud with the surveyed coordinate values and applying a transformation, the entire point cloud can be aligned to the local coordinate system. In photogrammetry using smartphones or drones, you can also include several known markers in the images and later perform coordinate correction in photogrammetry or image analysis software.

Ensuring high-precision coordinate control allows plan views to be correctly overlaid with other equipment drawings and topographic maps. In longitudinal profiles, you can also compare design elevations (for example, the height of a roadway subbase or the existing pipe invert elevation of a sewer) with the heights from the scanned point cloud, making discrepancies clear. If coordinate alignment is lax, even carefully derived drawings from point clouds will be offset from existing drawings and fail to gain trust, so enforce consistent management of positioning accuracy from acquisition through processing. In particular, public surveys and infrastructure drawings require conformity to Japan’s geodetic datum (JGD2011/2020) for coordinates and heights (geoid height). With strict referencing using RTK-GNSS and known control points, point cloud data can serve directly as reliable material for design and construction drawings.

Tip 3: Merging and Filtering Point Cloud Data

Point cloud data acquired on site contain noise and unnecessary areas as is, so they are preprocessed before being drafted. The first thing to consider is the integration (merging) of point clouds. Terrestrial laser scanners and smartphone LiDAR may not capture everything without blind spots in a single scan. For example, scanning only from one side of a trench may leave the opposite wall or the back side of pipes in shadow, resulting in no points. Therefore, it is necessary to perform multiple scans from different positions and angles and overlay the point clouds to fill gaps. For integration, in addition to the previously mentioned method using targets, algorithms such as ICP (Iterative Closest Point), which align point clouds using their characteristic shapes, are used. While matching overlapping points, deviations are corrected at the millimeter level (mm; 0.04 in) to produce a unified point cloud. Especially when long pipelines are measured in sections, check that the point clouds of each section are properly connected, and, if necessary, make error adjustments on the order of several centimeters (cm; a few in).

Next, we perform noise and unnecessary point removal (filtering). Point clouds from construction sites often capture elements unrelated to mapping buried pipes, such as people, heavy machinery and vehicles, temporary supports, and obstacles. In point cloud processing software, clearly unnecessary point-cloud regions are deleted to tidy the data. For example, isolated points floating in the air (points caused by laser misreflection) and obvious outliers (points that deviate significantly from their surroundings) can be removed with an automatic filter. Also, point clouds that are far above the ground surface (such as the tip of a crane or parts of nearby buildings) are unnecessary for buried-pipe drawings and can be deleted in bulk. In addition, fine noise layers caused by measurement errors can be removed with smoothing filters or by specifying ranges.

Trimming the area is also effective. By extracting only the necessary area around the buried pipe from the acquired point cloud data and working with that subset, you can reduce data volume and increase work efficiency. For example, even if you scanned widely, instead of the full width of the road or the entire site, you keep only the point cloud near the trench where the pipe is buried and thin out the rest. Point clouds of unrelated building interiors or distant terrain are also subject to deletion. Such filtering and trimming yield a clean dataset that contains only the point clouds necessary to represent the buried pipe. This serves as the basis for drafting work and smooths subsequent processes.

Tip 4: Classification and Extraction of Buried Pipes and Ground Surface

In point cloud processing, it is necessary to distinguish and handle the point cloud of the buried pipe—the object of interest—separately from other point clouds. Therefore, through a point cloud classification (classification) task, points are assigned to classes such as buried pipes, ground surface, other structures, and vegetation.

In typical 3D laser scan data, it is common to first extract the point cloud of the ground (ground surface). Because the ground is located above buried pipes and distributed over a wide area, automatic extraction by algorithms is relatively easy. For example, a low-point filter or a cross-section method can classify points with height differences above a certain threshold into the ground class. In the case of buried pipe construction, the idea is that the existing terrain and road surfaces spreading around the trench are extracted as the "ground surface class". At the same time, the point clouds of pipes and buried objects are identified. If excavation is in progress, the surface points of the pipe should be present in the point cloud as a band. For straight pipes they appear as cylindrical shapes, and at joints or bends they appear in the point cloud as curved or branching shapes.

In some cases, automatic recognition algorithms can be used. Recent point-cloud processing tools include functions that use machine learning to recognize specific shapes such as pipes and columns. By using these, the tools can detect cylindrical shapes in the point cloud from scan patterns and changes in point density, and extract them as buried-pipe candidates. However, in typical infrastructure sites, soil and other structural point clouds are mixed in the surroundings, so it is often difficult to extract only the pipes automatically and accurately. Therefore, a semi-automatic approach is practical: first roughly select the point cloud along the pipe route by hand, remove the unnecessary parts, and then designate the remainder as the pipe class.

In addition to buried pipes, classifying ancillary structures is also useful. If structures that connect to the pipes, such as manholes and valves, are included in the point cloud, mark those as separate classes as well. With point cloud data classified in this way, analysis becomes much easier—for example, displaying only the point cloud of the "pipe class" to examine pipe geometry, or combining it with the "ground surface class" to evaluate pipe depth. Once the point cloud of buried pipes has been clearly extracted through the classification work, you can move on to the stage of creating cross-sectional and longitudinal drawings based on that.

Tip 5: Identifying Pipe Positions by Extracting Cross-Sections

Creating cross-sectional views is indispensable for understanding the characteristics of buried-pipe point clouds and producing precise drawings. First, this section explains how to extract transverse sections (cross-sections) relative to the pipeline and how to utilize them. A cross-section is a view of the cut produced by a plane perpendicular to the pipeline’s longitudinal axis, and it is used to inspect the ground and the pipe’s cross-sectional shape.

To obtain cross-sections from point cloud data, specify arbitrary section lines in the analysis software. Typically, set the section plane perpendicular to the pipe centerline. For example, it is effective to cut sections every 10 m (32.8 ft), or at changes in terrain or the pipe (bends or slope change points). It is also important to set the width of the section plane (slice thickness) appropriately. If it is too thin, the point cloud will be sparse and hard to see; if it is too thick, points outside the section will be included and the shape will become unclear. In general, slice at a thickness of about half the pipe diameter up to about the same as the diameter, and display the points within that section by projection.

When you view the generated cross-section, you can immediately see the relationship between the ground surface and the buried pipes at that location. The point cloud of the ground surface appears on the cross-section as a ground line (topographic profile), depicting the shape of the trench and the undulations of the surrounding ground. The point cloud of the buried pipe is projected onto the section, and if the section passes near the pipe’s center, the pipe’s outline will appear almost circular. From this section, you can directly measure the elevations of the pipe crown and pipe invert, and the burial depth (the distance from the ground surface to the pipe crown). In addition, the pipe diameter can be confirmed by measuring the diameter on the section. When multiple buried objects run in parallel, the cross-section can show their relative positions (spacing distances and vertical differences).

Furthermore, cross-sectional drawings can also be used for comparison with the design sections. If standard design cross-sections or pipe layout drawings drawn before construction are available, they can be overlaid and compared with the as-built sections obtained from the point cloud. For example, if the design specified that a pipe be buried 1.5 m (4.9 ft) below the ground surface but the actual section shows it was 1.6 m (5.2 ft) below, the difference in burial depth can be quantified. In this way, capturing the as-built condition by extracting point cloud sections is a powerful tool for quality control and record-keeping. The extracted cross-sections can also be output as CAD data and turned into drawings. If the section lines are converted to DXF etc., they can be easily reused as figures attached to reports or as materials for maintenance and management. Creating multiple cross-sections does not significantly increase the workload as long as the point cloud is available, so the more finely they are cut along the length of the pipeline, the more detailed the record will be.

Tip 6: Generating Pipe Centerlines and Creating Longitudinal Profiles

Once you have captured the conditions at each location in the cross-sections, the next step is to create a longitudinal drawing for the entire piping. The longitudinal profile (profile drawing) is a sectional view sliced longitudinally along the buried pipe, and it continuously represents changes in the pipe’s gradient and depth. To create this, you must first determine the trajectory on the point cloud data that serves as the centerline of the buried pipe.

Specifically, we extract a series of points that pass through the center of the pipe from a point cloud classified as pipe. For straight piping, the point cloud is arranged in a straight line along the pipe axis, so picking a few points that form its center will generally determine the path. Even if there is a gradient, you can trace the shape by taking points on the top or bottom of the pipe at regular intervals. More rigorously, you can perform cylindrical fitting to the point cloud (fitting a cylinder using the least squares method) and compute its center coordinates. For example, by fitting a circle to the point cloud of a pipe with a diameter of 300 mm (11.81 in) to compute the center, and sliding it slightly while determining the center axes of consecutive cylinders, you can derive the three-dimensional route of the piping with high accuracy.

Once a series of coordinate points (X, Y, Z) that constitute the pipe centerline are obtained, connect them in sequence to create a three-dimensional polyline (polyline). To draw a longitudinal profile, represent this 3D centerline as a graph of elevation change versus horizontal distance. Specifically, calculate the cumulative distance from the start point on the centerline to each point (the so-called "station" or "chainage") and use that as the horizontal axis. Use the elevation at each location for the vertical axis. The resulting longitudinal profile becomes a curve with the horizontal axis showing distance along the pipe and the vertical axis showing elevation. To match the format of longitudinal sections in design drawings, include the ground surface profile as well if necessary. The ground surface profile can be obtained by plotting the ground directly above the pipe, if it is available as a point cloud, at the same chainages as the centerline.

From a longitudinal profile, the pipe’s slope, bends, and depth can be read continuously. For example, in the design longitudinal profile of a sewer pipe, the plan is to descend at a constant gradient (for example, 1/100), but by overlaying it with the as-measured longitudinal profile, you can check for construction errors or sag. Also, at intersections with other buried utilities or underground structures, if you indicate the elevation relationships at the intersection locations on the longitudinal profile, clearance checks become easy. Longitudinal profiles derived from point clouds capture much finer variations than manual surveying. Slight vertical deviations that occur during construction, as well as the positions and elevations of fittings and valves installed in between, can be shown accurately on the profile. Such detailed longitudinal profiles are not only valuable as as-built documentation, but they also serve as an information source during future excavation planning, indicating “a buried pipe previously runs at a depth of ○ m (○ ft) here.”

Tip 7: Key points for creating floor plans from point clouds

Alongside longitudinal profiles, creating plan views of buried pipes is also one of the goals of point cloud utilization. A plan view is a drawing that shows the piping route as seen directly from above, allowing you to understand the paths of lateral piping, bends, and outlet positions. To generate a plan view from a point cloud, you basically extract the horizontal position information of the buried pipes and convert it into two dimensions.

If you have the pipe centerline data obtained in Tip 6, you can draw the pipe’s plan-view route simply by projecting it onto the XY plane. By extracting the X and Y coordinates of each point of the 3D centerline and connecting them in order, you directly obtain the pipe line on the plan. When finishing the drawing, you may smooth and refine the geometry by fitting arcs or straight segments at appropriate corner points (bend points), but fundamentally it is fine to remain faithful to the measured point-cloud values. If the pipe is nearly straight, its direction angle and length will be obtained, and bent pipes or curves will appear in their actual shapes following the point cloud.

When creating a plan drawing from a point cloud, capturing surrounding reference elements as well makes the drawing easier to understand. For example, when drawing over a road, if you trace the road boundary and curb line from the point cloud and draw them in thin lines, you can immediately tell which lane a pipe lies under. For a site, if you have captured building exterior walls, boundary lines, and tree locations in the point cloud, you can incorporate them into the plan view as well. For tracing such features, using the nadir view (ortho-projected image) of the point cloud is convenient. Generate an image in point cloud processing software that flattens each point as seen from directly above, and use that as an underlay in CAD to trace pipe routes and the outlines of structures. You can create accurate plan drawings with a feeling much like tracing a high-resolution aerial photograph.

The completed plan becomes a record drawing showing the location of buried pipes, paired with the longitudinal profile. If pipe diameter, material, burial date, and other details are added to the drawing, its future value as a reference will be further enhanced. It is effective to manage the created plan data (CAD data) in conjunction with the 3D point cloud model. Details that cannot be fully represented on a 2D drawing can be confirmed by viewing the original point cloud model, and if there are questions about the plan, you can return to the point cloud for reanalysis. By linking the point cloud and the plan, you will be able to respond quickly to future reverse engineering (as-built modeling) and the design of additional works.

Tip 8: Improving Efficiency by Leveraging Smartphone LiDAR + GNSS

As a final point, I will mention that by using recently introduced smartphone-integrated point cloud measurement tools, the drafting work for buried pipelines can be made dramatically more efficient. Point cloud measurement, which traditionally required high-performance laser scanners and surveying instruments, can now increasingly be substituted by a combination of a smartphone and a small GNSS receiver. For example, using an RTK-GNSS-capable device that can be attached to smartphones such as the iPhone together with a dedicated app, a high-precision 3D scan can be completed on site. Specifically, the system scans pipes and terrain inside a trench with the smartphone’s built-in LiDAR sensor while the attached GNSS receiver measures positions with centimeter accuracy (cm level accuracy, half-inch accuracy) and integrates them in real time. Because point cloud acquisition and conversion to geodetic coordinates proceed simultaneously, simply walking around the site while scanning yields point cloud data with absolute coordinates immediately.

The advantages of such smartphone surveying systems lie primarily in mobility and ease of use. Because field technicians can perform surveys with a single smartphone without bringing heavy tripods or large equipment, the hurdles for personnel allocation and equipment preparation are reduced. In piping work, sudden design changes or additional tasks can occur, but even then it becomes easy to quickly perform additional scans on site and update plan and longitudinal profile drawings. Real-time processing and cloud integration are also major strengths. Point clouds captured by the smartphone are automatically uploaded to cloud services, and some systems perform AI-based noise removal and automatic alignment of the point clouds. For example, a 3D model can be generated in the cloud within minutes of capture, allowing immediate review from an office PC. This makes it possible to instantly check on site for any omissions and, if anything is missing, immediately take additional measurements for on-the-spot verification.

The use of smartphone LiDAR plus GNSS for buried-pipe surveying is also bringing a transformation to the creation of as-built drawings. For example, by using an iPhone-mounted high-precision GNSS positioning device like LRTK, anyone can record point clouds of buried pipes and produce drawings in a short time. With solutions such as LRTK, you simply scan the piping and upload it to the cloud, where the pipe’s shape and depth are automatically processed, and auxiliary functions can generate longitudinal and cross-sectional diagrams as needed. Furthermore, the point cloud data stored in the cloud can be utilized for future maintenance. Combined with a smartphone’s AR capabilities, you can display the buried pipe on-site in a see-through view and intuitively visualize that “a pipe runs ○ m (○ ft) below ground at this point.” The task of locating underground installations, which traditionally relied on paper drawings and buried markers, is becoming an era in which anyone can accurately perform it using point clouds + RTK + AR.

Overall, leveraging cutting-edge technologies that combine smartphone LiDAR and RTK-GNSS greatly streamlines the process of creating plan and profile drawings for buried pipelines and improves data accuracy. Because the entire workflow—from point cloud acquisition to drafting—can be completed by a single operator with a single device, internalizing surveying and drafting work that was previously outsourced to specialist companies is no longer out of reach. Leaving reliable and detailed records in buried pipeline construction contributes to safer infrastructure maintenance and reduced future construction risks. Be sure to adopt the latest tools suited to your site, take into account the eight tips introduced, and use point cloud data to help produce high-quality drawings of buried pipelines.