Easy even for beginners! How to export cross-sections from point clouds to DXF

Table of Contents

• Introduction

• Basic knowledge of point cloud data and cross-section drawings

• Step 1: Acquire and prepare point cloud data

• Step 2: Extract cross-sections using point cloud processing software

• Step 3: Export cross-section drawings to a DXF file

• Step 4: Use the exported cross-section drawings in CAD

• Tips and precautions for creating cross-section drawings

• Recommendation for simplified surveying with LRTK (Conclusion)

• FAQ

Introduction

Methods for creating required cross-sections from three-dimensional point cloud data and exporting them in the general-purpose CAD format DXF have been attracting increasing attention in recent years. Until now, leveraging survey data and creating drawings has tended to be seen as a specialized and difficult task. However, even beginners can easily obtain cross-sections from point clouds and export them to DXF. In this article, we will clearly explain the concrete procedures and key points. As digital surveying and DX promotion at construction sites advance, the skill of exporting cross-sections from point clouds to DXF will greatly contribute to improved operational efficiency and accuracy. First, we will look at the basic knowledge in order.

Basic Knowledge of Point Cloud Data and Cross-sections

Point cloud data refers to a collection of numerous points acquired from objects or terrain using laser scanners or photogrammetry (photogrammetry). Each point contains coordinates (X, Y, Z) and color information, and the aggregation of points represents the shape of the object in three dimensions. Examples include point clouds obtained by laser scanning the interior of a building and terrain point clouds generated from images captured by drones. In recent years, it has also become possible to easily acquire point clouds using smartphones equipped with LiDAR functionality.

Meanwhile, a cross-section is a drawing that represents the internal shape (profile) of an object such as a building or terrain when it is cut vertically along a given cutting plane. In civil engineering, it is used for longitudinal and cross-sectional views of terrain, and in architecture for building sections; these are widely employed in design and construction. By creating cross-sections from point cloud data, the current on-site shape can be accurately represented as a two-dimensional drawing, aiding comparison with design drawings and verification of the as-built (final shape).

DXF format (Drawing Exchange Format) is a drawing data exchange format supported by many CAD applications. By exporting cross-sectional drawings in DXF, you can import and use the section lines in various CAD and surveying software. In other words, DXF export is the bridge from "3D point cloud data" to "2D drawing data that anyone can handle." Below, we introduce the concrete steps to extract cross-sections from point clouds and save them as DXF files.

Step 1: Acquire and Prepare Point Cloud Data

First, prepare the source point cloud data. If you already have point clouds acquired with a scanner or a drone, import them into your PC. If you are going to acquire them now, in addition to high-precision terrestrial laser scanners and drone photogrammetry tools, smartphone LiDAR can also be used for small-scale point cloud surveys. For beginners, accessible methods include generating point clouds from photos using free photogrammetry software or loading publicly available data into open-source 3D point cloud processing software.

Point cloud data comes in various formats, but commonly formats such as `.las`, `.ply`, and `.xyz` are used. Once you have acquired the data, import it into point cloud processing software on your analysis PC. At this stage, you need to pay attention to the settings for the coordinate system and units. For example, if measurements were taken in the Geospatial Information Authority of Japan standard coordinate systems (such as the plane rectangular coordinate system or geographic coordinates of the World Geodetic System), specify the correct coordinate system in the processing software to ensure consistency with later drawings. Also, extracting only the required parts from the entire point cloud and removing or thinning noisy points (to reduce data volume) will make subsequent steps smoother.

As an additional preparatory step, consider in advance the position from which to extract the cross section. Decide which line you want the cross section for (for example, cross-sections at each survey point along a road, or particular sections of a building), and by identifying the position and orientation of the corresponding cutting plane, subsequent work will be clearer.

Step 2: Extract cross-sections using point cloud processing software

Next, perform the cross-section extraction on software capable of point cloud processing. I will not name specific professional software here, but many point cloud processing tools and some CAD programs include cross-section extraction features. The basic operation steps are as follows.

• Load the point cloud data: Open the point cloud prepared in Step 1 in your point cloud processing software. For high-precision measurement data, the number of points can be on the order of several million to several tens of millions of points, but some software simplifies the display (sampling display) to make it easier to handle. Beginners should first check the overall view and try zooming into the area where they want to take a cross-section.

• Setting the section plane (section line): Use functions in the software such as "section" or "slice" to set a plane that cuts the point cloud. For example, for a vertical section you can place a plane perpendicular to a chosen X or Y coordinate position to extract the points that intersect that plane. Also, for sections along an arbitrary line (such as cross-sections along a curved road), some software lets you specify a guiding segment or polyline to obtain the terrain cross-section directly beneath it. Slice thickness (thickness): if you can set this parameter, adjust the section thickness appropriately. If a thickness of 0 (cutting as a plane) causes the points to be sparse, you can slice into a band on the order of a few centimeters to a few tens of centimeters (a few in to a few tens of in) to gather more points and make it easier to draw lines later.

• Extraction of the cross-section point cloud: When you cut the point cloud at the specified plane position, only the points that lie on that cross-section are extracted. In the software these will appear as a new point cloud object (cross-section point cloud), or only the points on that cross-section will be highlighted in the view. Verify that the extraction results match the intended area. For example, if you wanted a cross-section of the ground surface but points from trees overhead are included, additional processing will be required to remove them. Delete unwanted points using filters or manual selection, leaving only the point cloud that defines the cross-section.

• Generation of cross-section lines (vectorization): Create a cross-section line (polyline) from the point cloud on the extracted cross-section. Many tools have functions for drawing lines from points (e.g., "polyline drawing" or "spline curve fitting"). Some provide batch automatic approximation to a line, while others offer tools to manually draw lines by tracing the points. For beginners, it is recommended to use software with an automatic polyline generation feature. Even when using automatic features, check that the line is drawn smoothly and manually correct parts as necessary. Note: Where points are missing or discontinuous, the line may be fragmented or take on unnatural shapes. In such cases, better results can be achieved by connecting the line while interpolating using surrounding points.

Although the interface and terminology vary by software, you can generally follow the above workflow to extract cross-sections (line data) from point clouds. It may involve trial and error at first, but it's a good idea to practice with small datasets to get the hang of it.

Step 3: Export the cross-section drawing to a DXF file

Once you have the line data of the cross-section, it's finally time to export it in DXF format (write out). DXF output can usually be selected from the "Export" or "Save As" functions in many point cloud processing or CAD software. Proceed while paying attention to the following points.

• Selection of export target: Select the cross-section polyline or cross-section point cloud generated in the previous step and save it in DXF format. Because some software exports everything that is currently displayed while others export only selected objects, it is safest to copy just the cross-section lines into a new file before exporting if necessary.

• Coordinate system and unit check: Verify that the numeric coordinates exported to the DXF file are correctly recorded in the original data's coordinate system and unit system. For example, if you created a cross section from a point cloud measured in meter-based coordinates, the DXF will also use meter units (m (ft)) for its coordinate values. Some commercial point cloud software embeds coordinate system information into the DXF as metadata, but because not all software does this, make sure you know which coordinate system is being used. This is important to prevent having to perform coordinate transformations later on the CAD side.

• Layer and color settings: Because the DXF format has a layer feature, you can assign lines to specific layer names when exporting. If possible, exporting with an easily understood layer name such as "section view" makes it convenient to toggle visibility by layer later in your CAD software. You may also be able to set line colors and linetypes. By default it tends to be a single color (for example white/black), but you can color-code as needed. That said, because colors and linetypes can be edited within CAD, simple settings at the time of export are acceptable.

• File name selection: Give the DXF file you save a clear, easy-to-understand name. For example, using a name that indicates the site and cross-section location, such as "SiteA_CrossSection1.dxf", will reduce confusion later when handling multiple cross-section files.

Confirm the above settings and export the DXF file. The processing time depends on the complexity of the section lines, but it usually does not take much time. Because DXF files are text-based, they are relatively lightweight and easy to share as email attachments. Once the created DXF section drawing has been correctly output, try importing it into CAD software in the next step.

Step 4: Utilize the exported cross-section drawings in CAD

Once you export the DXF file, use the cross-section data in CAD software. The procedure for importing DXF into general-purpose CAD tools and civil engineering design software is largely the same. Below is an explanation of the typical workflow.

• Loading the DXF file: Start your CAD software, choose "Open" or "Import" from the "File" menu, and specify the DXF file you saved in step 3 (some software allows you to load a DXF simply by dragging and dropping it). If the import options ask for the unit system, choose the same units used when acquiring the point cloud data (meters (ft) or feet, etc.).

• Check display of the cross-section drawing: Confirm that the loaded DXF cross-section drawing is displayed correctly in the CAD. If the section lines are displayed very small, conversely too large, or appear to be off-screen and not visible, adjust the zoom or display range settings. Common occurrence: With data that has large absolute coordinate values, such as survey coordinate systems, lines may be located far from the origin when opened in CAD and become hard to find. In that case, using the fit-to-screen command (e.g., "Fit to Screen" or "Zoom Extents") will capture the cross-section drawing.

• Adjustment of coordinate positions: Check whether the lines of the imported section drawing are located where you expect them to be on the CAD drawing. For example, if the original remains in field coordinates, it may not align with other drawing elements within the CAD. In that case, use a reference point (for example, a known point on the section line) to align with the other drawings (translation and rotation). If the point cloud was processed in the same coordinate system as the design drawings, this alignment step is unnecessary, and the section DXF should directly overlap in the correct position.

• Editing as a drawing: Cross-section drawing data are usually handled in CAD as polylines or spline curves. If necessary, you can retrace these lines to smooth them or annotate height information. If you want to calculate the cross-sectional area, you can also enclose it with a polyline and measure the area. Also, if layer settings were configured at export, you can change line color and line type per layer on the CAD side. When loading multiple cross sections for comparison display, visibility is improved if they are separated into layers.

Section drawings imported into CAD software in this way can be used in combination with other design elements. For example, you can overlay them with the planned section lines from the design stage to check discrepancies with current conditions, or compare as-built results after construction with the design sections to calculate cut-and-fill volumes. Importing via DXF allows point-cloud-derived data to be treated in the same way as conventional 2D drawings, making it easy to integrate smoothly into workflows, which is a major advantage.

Tips and Precautions for Creating Cross-Sectional Drawings

There are several tips and points to keep in mind to successfully create cross-sections from point clouds and export them to DXF.

• Thin out large point clouds appropriately: If the original data's point cloud is too dense, unnecessary points may be included during section extraction, causing the lines to become jagged. Either pre-select only the point cloud near the target section and reduce (sampling) it, or delete points that are outliers after extraction. If points are too densely clustered, software display and processing will also become heavy, so it's important to keep a moderate density.

• Smoothing cross-section lines: If the extracted cross-section lines look uneven, consider smoothing (refinement) or converting them to spline curves in CAD. If the original point cloud has noise, the lines tend to be irregular, so shaping them into smooth lines with the idea of taking representative values makes them easier to read. However, to avoid removing important features such as subtle terrain undulations, it's best to make adjustments while comparing them with the original data.

• Understanding the height reference: Section drawings include information in the vertical direction (Z-axis). For civil engineering cross-sections, they often show elevations relative to a reference elevation; for architectural sections, they often indicate heights from the 1st floor or ground level (GL). If the point cloud was surveyed and tied to known on-site elevation points, there is no problem; otherwise, you will need to adjust a vertical offset afterward. After importing the DXF section into CAD, it is also possible to move the entire model up or down to match the height reference.

• Managing multiple cross sections: When a single project requires multiple cross-sectional drawings, it is efficient to use functions in point cloud processing software that slice at regular intervals or tools that bulk-generate a series of cross-section lines. For example, in cases where cross sections are extracted along a road at 10 m (32.8 ft) pitch, you can also use automatic batch processing to output DXF files all at once. If creating them manually one by one, manage by adding sequential numbers or station numbers to file names to make organization easier.

• Data integration trials: The procedure for exporting cross-sections from point clouds to DXF may not go perfectly on the first try. Problems can occur, such as positions not matching due to coordinate system shifts or lines not being output correctly. At first, it is safer to test on a small area and confirm import into CAD before running with production data. If problems arise, isolate which stage the cause is in (for example, a misconfiguration in the point-cloud processing software or the CAD import settings) and address it.

If you pay attention to the points above, even beginners should be able to reliably convert point cloud cross-section drawings to DXF. At first, proceed cautiously and check your work as you go; as you gain experience, you'll be able to process many cross-sections more efficiently.

Recommendation for Simple Surveying with LRTK (Final Note)

Finally, together with creating cross-sections from point clouds, we introduce a new simple surveying method you should know. Traditionally, small on-site measurements and checks were often handled with convenient tools such as tape measures and handheld GPS, forcing compromises in accuracy. However, today it is possible to use a solution called LRTK to "dramatically improve accuracy while keeping the same ease of use."

LRTK is a cutting-edge surveying tool that combines an ultra-compact, high-precision GNSS positioning device that can be mounted on a smartphone with the smartphone’s camera and LiDAR capabilities. Without using specialized heavy equipment or special scanners, you can perform centimeter-level positioning and point cloud measurement with just a smartphone (cm level accuracy, half-inch accuracy), dramatically streamlining on-site simple surveying. For example, rather than laser-scanning a wide area to create cross-sections, you can use LRTK to directly measure the necessary points on-site (such as key elevations and widths of the terrain) and grasp the profile on the spot.

By introducing LRTK on-site like this, on-site staff themselves can quickly handle some tasks that previously relied on specialized surveying departments or outsourcing. Being able to check cross-sections and verify heights immediately, without waiting for surveying, keeps construction progress management running smoothly. Even in situations that demand high accuracy, you can confidently entrust simple surveying to LRTK.

As high-precision positioning becomes easy to use, democratization of surveying will advance, including the acquisition of point cloud data and the creation of cross-sectional drawings. LRTK, which can be operated even by beginners without specialized knowledge, is truly a "surveying instrument that fits in your pocket." Together with the digital-data utilization skills cultivated through converting point clouds into drawings, please take on next-generation smart surveying. As a reliable ally supporting on-site DX (digital transformation), LRTK will help drive improvements in your operations.

FAQ

Q: Is expensive software necessary to create cross-sections from point clouds? A: You don't necessarily need to purchase expensive specialized software. Open-source point cloud processing software and some free applications can extract cross-sections and output DXF. There are also cloud-based services that process point cloud data and generate cross-sections. Paid professional software offers advanced features, but for beginners who only need basic cross-sections, free tools are often sufficient.

Q: Can I obtain point cloud data without owning a laser scanner? A: Yes, there are ways to obtain point cloud data without a laser scanner. A common method is photogrammetry, which uses software to generate point clouds from many photos taken with a digital camera or a drone. Recent smartphones also include LiDAR scanners in some models, allowing you to easily scan the surrounding environment as point clouds using an app. You can also use existing open data (for example, point clouds included in the Geospatial Information Authority of Japan's Basic Map Information). Furthermore, using a smartphone-linked surveying device like LRTK makes it easy to perform high-accuracy point cloud measurements yourself.

Q: Can cross-section processing be performed on a regular PC even with large point cloud data? A: Point cloud data tend to be large, but with some measures, cross-section processing is possible on a typical PC. The key is to work with only the necessary area. Trying to process the entire massive point cloud at once makes you want a high-performance PC, but only a portion of the point cloud is used for cross-sections. If you crop out only the points near the target cross-section within the software or downsample to reduce data volume, a PC with standard specifications will run more smoothly. Also, leveraging cloud services to process point clouds on a server and then downloading only the resulting cross-section data is an effective approach.

Q: When I import a cross-sectional drawing exported as DXF into CAD, the position is wrong? A: First, check for mismatches in coordinate systems and unit systems. For example, if the point cloud was recorded in a local coordinate system with its own origin, it may not align with other drawings in CAD. In that case, you will need to manually align it using a common reference point. Also, an incorrect unit setting on import can cause a scale error (for example, data in meters (m, ft) being treated as millimeters (mm, in), causing a 1000× offset). DXF files often do not contain explicit unit information, so make sure to specify the units of the original data correctly in CAD. If it still does not align, reconfirm the coordinate system used during the original point-cloud processing and, if necessary, apply the appropriate coordinate transformation before re-exporting.

Q: What should I do if the extracted cross-section line is jagged and hard to read? A: A jagged cross-section line is usually caused by the coarseness or noise in the point cloud. Remedies include smoothing and interpolation. Smoothing means applying filters in CAD to make polylines closer to curves, or removing unnecessary kinks while retaining key points. Also, if parts of the point cloud are missing in a terrain cross-section and the line is interrupted, it's necessary to interpolate that segment following the surrounding slope (manually adding line segments). If extreme outlier points are present in the source data, removing them at the cross-section extraction stage will make the line considerably smoother. In short, you can produce a clear cross-section by combining cleanup of the original data + post-processing of the line.