No coordinate offset! High-precision point cloud to DXF conversion and data extraction with LRTK

Table of Contents

• Introduction

• What are point cloud data and DXF

• Background: why point cloud to DXF conversion is needed

• Coordinate offset issues arising from point cloud → DXF conversion

• Key points to prevent coordinate offsets

• Procedure to convert point cloud data to DXF

• High-precision, offset-free DXF conversion with LRTK

• Recommendation: simple surveying with LRTK

• FAQ

Introduction

Have you ever obtained high-precision point cloud data and then struggled because the positions didn’t match when imported into CAD drawings? Even if measurements are taken at millimeter-level accuracy, if the point cloud is not displayed at the correct location on the drawing due to coordinate offsets, it can lead to serious mistakes and rework. To prevent such “coordinate troubles,” it is important to correctly understand the differences between the coordinate system used by the point cloud data and the coordinate system used by the design drawings, and to take appropriate measures. This article explains the common causes of coordinate offsets when exporting and sharing point cloud data in DXF format, and how to prevent them. We organize and present key points so that those involved in surveying and design can exchange data with confidence.

What are point cloud data and DXF

Point cloud data are digital data that record a large number of points forming the surface of an object such as a building or terrain as a set of three-dimensional coordinates (X, Y, Z values). They can be obtained by laser scanners, photogrammetry, or, more recently, LiDAR-equipped smartphones, and they provide precise representations of object shapes. Point clouds are characterized by vast collections of coordinates—ranging from millions to hundreds of millions of points—that can capture subtle surface irregularities of walls and ground.

DXF (Drawing Exchange Format), on the other hand, is a file format widely used for CAD data exchange. It stores information such as lines, points, and shapes that represent drawings in a text format, and is commonly used to transfer data between different CAD software. In construction and civil engineering, DXF files are often specified for exchanging design drawings and survey data. DXF is a flexible format that can store not only 2D drawings but also 3D coordinate data, and its software-independence is a major advantage.

Background: why point cloud to DXF conversion is needed

Converting point cloud data to DXF is mainly done to utilize point cloud information in design and construction workflows. From detailed point clouds obtained by laser scanning or photogrammetry, there are increasing cases in which dimensions in required locations are measured or plan and section drawings are generated. However, point cloud data themselves are large and cumbersome to handle, so simplification or vectorization is often necessary before directly incorporating them into design drawings. For example, you might trace building outlines from point clouds into DXF drawings, or extract terrain section lines for civil design. Once converted to DXF, those vector drawings can be edited and annotated in commercial CAD software, allowing point cloud-derived information to be smoothly integrated into the design flow.

Converting point clouds to DXF is also useful when handing data to a third party. Even if the recipient does not have a dedicated point cloud viewer or processing software, a DXF drawing can be opened in common CAD software for review. In short, by extracting the essence of point cloud data into DXF format, data sharing and secondary use become much easier.

Coordinate offset issues arising from point cloud → DXF conversion

The biggest issue to watch for when converting point clouds to DXF and importing them into drawings is coordinate offsets. If the coordinate values attached to the point cloud differ from the coordinate reference used by the design drawing, the numerical values for the same location will not match and the plotted positions will differ on the drawing. Here are several specific possible causes.

• Difference in coordinate systems: If the coordinates assigned when acquiring the point cloud differ from the coordinate system used by the design drawing, large horizontal position offsets can occur. For example, if a point cloud obtained by drone or smartphone has global geodetic coordinates (earth-referenced latitude/longitude or absolute coordinates), while the drawing is expressed in a local coordinate system that sets a site-specific point as the origin, the numerical values will differ greatly even though they refer to the same object. Naturally, they will not overlap as-is and the point cloud will be displayed in a completely different location. Even when both use nationally defined public coordinate systems (such as plane rectangular coordinate systems), a difference in zone number will change the origin and can produce offsets of hundreds of kilometers.

• Difference in geodetic datum: Older drawing data may have coordinates recorded in an older datum rather than the currently used global geodetic datum. In Japan, the official standard switched to the global geodetic system (JGD2000/2011) in 2002, but earlier data may use the old Tokyo Datum. There is a persistent offset of several hundred meters between the old datum and the current global datum depending on the region, so overlaying old and new data without conversion will not align positions.

• Difference in unit systems: Don’t overlook differences in coordinate units such as meters (m) versus millimeters (mm). For example, if coordinates on a drawing are “12000, 5000” and the unit is mm, converting to meters yields 12.000 m (39.370 ft), 5.000 m (16.404 ft). Misinterpreting units can result in a 1000× numerical error, causing large apparent offsets. When handling CAD data, pay careful attention to the drawing’s unit settings.

• Lack of localization (no coordinate fitting): Transforming a point cloud with absolute coordinates from GNSS to a drawing that uses a site-specific local coordinate system is called localization (coordinate fitting). If you overlay data without performing coordinate transformation based on known control points, you will obviously get discrepancies of tens of meters or more relative to the site plan. For drawings using local coordinates, it is essential to compute offsets and rotation angles from at least 2–3 known points and apply corrections to the survey coordinates. Failing to do so will cause the point cloud to be plotted in the wrong location on the drawing.

• Axis tilt and scale differences: Even if a single point is matched, if the site coordinate axes are rotated relative to true north or a scale factor differs, the farther from the matched point you go the larger the offset becomes. When handling large-area point clouds or drawings whose XY axes are tilted from the north reference, take care. Unless you match multiple points to correct angle and scale differences, locations near the edges of the area can show significant position offsets.

Due to these factors, point cloud data you obtained may cause coordinate mismatches when overlaid on CAD drawings. So how can you prevent them? The next section explains key points.

Key points to prevent coordinate offsets

To ensure point cloud data and design drawings share the same coordinates, keep the following points in mind.

• Confirm the design coordinate system in advance: Before starting point cloud measurement or surveying, confirm the coordinate system to be used for the project. Identify whether the drawing uses a global geodetic system (e.g., JGD2011 plane rectangular coordinate system zone X), an old datum, or a local coordinate system with a specific point as the origin. At the same time, confirm the drawing data’s coordinate units (m or mm) and the vertical reference (e.g., sea level or another reference). Share this information with the client and design personnel in advance to unify recognition on site.

• Set the coordinate system for positioning data: When acquiring point cloud or survey data with RTK-GNSS, etc., configure the receiver or app to output in the same coordinate system as the design drawing whenever possible. For example, when using Japan’s network RTK, select the appropriate plane rectangular coordinate zone number in the receiver settings; if the old datum is needed, apply conversion parameters beforehand. Also, if you set up your own base station, be sure to enter the base station’s reference coordinates accurately using official public coordinates. Acquiring data in a system close to the design coordinate system from the start significantly reduces later correction work.

• Localization using known points (coordinate fitting): If known control points or boundary stakes with known coordinates exist on site, make sure to use them. Include two or more (ideally three or more) known points within the point cloud dataset, and compare their coordinates with the drawing values to calculate offsets and rotation angles. Correcting the entire point cloud based on these calculations converts it to the local coordinate system. Matching against multiple points also allows correction of subtle scale differences so distant parts of a large site align without offset. When overlaying point clouds on drawings created in local coordinates, localization is the decisive step to prevent coordinate offsets.

• Adjust vertical reference: In addition to horizontal position, confirm the vertical (elevation) reference. If the Z values in the point cloud are GPS-derived ellipsoid heights, they will not match design drawing elevations (such as sea level-based heights). Measure one known height point such as a local benchmark, compute the difference between that point’s height in the point cloud and the drawing (the local geoid separation), and apply a uniform correction to the other point cloud heights. Modern GNSS receivers and apps often include built-in geoid models to automatically convert to orthometric heights, so make use of these features.

• Validate the results: After coordinate transformation or localization, verify the results using separate known control points. If discrepancies of more than several centimeters appear, recheck the precision of the control points and the correction calculations. If validation passes, you can be confident that point cloud and survey points obtained under that coordinate system are consistent with the design drawing at high accuracy.

• Share information among stakeholders: Make sure not only surveyors but also design and construction teams share which coordinate standard is used for the project. Document and disseminate decisions such as “origin at control point XX with east direction as X-axis” or “adopting JGD2011 plane rectangular coordinate zone YY” so that handing data to other companies or departments does not cause confusion.

If you follow the points above, you can greatly reduce the risk of coordinate offsets when importing point cloud data into CAD drawings. In particular, localization using known points requires some effort but is indispensable when considering later error prevention and correction workload.

Procedure to convert point cloud data to DXF

Now let’s check a typical workflow to convert point cloud data to DXF and deliver it to a recipient. As noted earlier, proceeding after properly aligning coordinate systems is a precondition. The following is an example workflow.

• Point cloud acquisition and preparation: Obtain point clouds by laser scanner, LiDAR-equipped smartphone, drone photogrammetry, or other methods. After acquisition, perform registration of multiple point clouds, noise removal, and removal of unnecessary parts as needed. For large point clouds, split the area into manageable tiles or prepare reduced-resolution sample point clouds. Adjust the coordinate system so that it matches the design drawing as described earlier. For example, when aligning to a local coordinate system, complete processing that translates and rotates the entire point cloud using known points.

• Create vector data from the point cloud: Although it is technically possible to directly convert raw point clouds to DXF, converting millions of points into a drawing makes the data heavy and difficult to handle. Therefore, it is common to extract the information you need from the point cloud and vectorize it. Using dedicated point cloud processing software or CAD software, display the point cloud and trace the necessary points and contours to convert them into polylines or survey points. For buildings, trace wall and column centerlines to draw floor plans; for terrain, extract cross-section or contour lines at set intervals. This process generates the line data for DXF drawings from the vast point cloud in a human-readable form.

• Import into CAD and placement: Import the created coordinate lists and line data into CAD software. Many CAD packages allow batch placement of points from a text-form coordinate list, or can read point cloud files (e.g., LAS) as external references and draft over them. If the point cloud processing software can directly export DXF vector drawings, you can open that in CAD. In any case, do not change coordinate values during import. Avoid adding offsets or changing scale; set the import options to faithfully reproduce the original coordinates. If the point cloud has been adjusted to the design coordinate system, the data should display largely overlapping existing drawing elements at this stage.

• Check units and origin: Once data are placed in CAD, confirm the drawing units and origin position. If the CAD drawing units are set to mm while the point cloud uses meters, the coordinates will be displayed at 1000× the intended scale. Change the drawing units to meters or apply unit conversion options during import as necessary. Also check that the drawing origin (0,0) is not unintentionally shifted. If coordinate alignment was done correctly, one of the known control points should plot at the same coordinates as the reference point on the drawing.

• Export to DXF: After placing and verifying the data, save the completed drawing in DXF format. The recipient may request DWG depending on their software, but DXF is a safe default when unsure. Before saving, organize layer names, point/line colors, and other attributes to make the drawing easy to read. The generated DXF file can then be shared so that the recipient can work with point cloud–derived drawing information in their CAD environment.

This is the general flow for converting point clouds to DXF. The key points are to align coordinate systems and units, and to extract only the necessary information to reduce file size while preserving the precision of the point cloud.

High-precision, offset-free DXF conversion with LRTK

As described above, careful handling and conversion of coordinate systems are required to utilize point cloud data, but recently solutions have emerged that perform these tasks semi-automatically and simply. One example is point cloud measurement with the LRTK series. Using LRTK, you can obtain high-precision point cloud data without worrying about coordinate offsets from the start and efficiently proceed through to DXF drafting.

LRTK is a high-precision positioning system that works with smartphones and is powerful for on-site point cloud measurement and surveying. By attaching a dedicated ultra-compact RTK-GNSS receiver to a smartphone, you can obtain centimeter-level position coordinates in real time—something ordinary GPS cannot provide. Combined with a smartphone’s built-in LiDAR scanner or camera photogrammetry, LRTK is an innovative system that allows anyone to easily measure high-precision 3D point clouds. The convenience of being able to carry a smartphone plus LRTK and quickly scan a site alone is a major benefit, eliminating the need for expensive, large surveying equipment.

In the LRTK smartphone app, you can flexibly set the coordinate reference system. You can select regional public coordinate systems (such as plane rectangular coordinates of the global geodetic system) or arbitrary local coordinate systems. If you input known on-site control point coordinates, the app can perform localization (conversion to the site coordinate system) with a single tap, so you don’t need to worry about complex calculations—the coordinate fitting completes in seconds. As a result, you can acquire point cloud data from the outset using the same reference as the design drawing, greatly reducing subsequent transformation and correction efforts. On actual construction sites, users report that by measuring 2–3 control points and setting localization in advance with LRTK, all subsequently acquired point clouds and survey points are recorded in the public coordinate system (i.e., the same coordinates as the design drawing). The obtained data can be applied directly to finalized CAD drawings without coordinate conversion errors, enabling a seamless transition to design work.

Another advantage of LRTK is automatic cloud storage and sharing of acquired point clouds and survey data. Once measurements are completed on the smartphone, data are immediately uploaded to a dedicated cloud where office PCs can access the point clouds via a browser. A cloud 3D viewer can display heavy point cloud data smoothly without relying on the local PC’s specs, and you can measure distances and areas or draw lines on the point cloud with the mouse to create plan drawings. The drawn vector lines can be exported on the spot as DXF files for further editing and dimensioning in commercial CAD software. The raw point cloud can also be downloaded in LAS or XYZ format as needed, making integration with existing BIM/CAD software straightforward.

Thus, by leveraging LRTK, you can measure point cloud data and proceed to drafting without worrying about coordinate offsets. The system is intuitive even for those without surveying or CAD expertise, and it also shortens the work time for experienced technicians. Users who introduced LRTK report feedback such as “It’s overwhelmingly easier than traditional surveying instruments yet highly accurate,” and “Carrying a smartphone per person makes site response much faster.” LRTK is a reliable partner that supports processes from field surveys to drawing creation.

Recommendation: simple surveying with LRTK

When implementing a new workflow that uses point clouds and DXF drawings, LRTK is a very promising solution. Combining high-precision GNSS positioning technology with cloud services, the LRTK series dramatically improves on-site surveying accuracy and operational efficiency. It significantly reduces the effort required for coordinate fitting and point cloud processing, and enables anyone to easily acquire and share accurate survey data.

In the construction and civil engineering industries, the Ministry of Land, Infrastructure, Transport and Tourism–led i-Construction initiative calls for digitalization from surveying through design and construction. The LRTK series is a modern tool supporting this trend and will strongly promote on-site digital transformation. By allowing site personnel—not just specialists—to perform tasks formerly left entirely to experts, overall productivity and quality assurance can be improved.

To meet the growing needs for high-precision surveying and data sharing, consider introducing LRTK. The era has arrived in which a smartphone plus LRTK can handle surveying through drafting in one continuous workflow. For the latest information and product details, see the LRTK official site: https://lrtk.lefixea.com/. If you are interested, please take a look. Using LRTK could evolve your company’s field operations to the next stage.

FAQ

Q1. How do I fit a point cloud with a different coordinate system to a design drawing? A. The basic approach is to “convert to a common reference.” First confirm the coordinate system adopted by the drawing, and transform the point cloud coordinates accordingly. Specifically, perform localization by applying offsets (translation), rotation, and scale corrections to the point cloud based on known control points. Use two or more known points to compute horizontal translation and rotation, and correct scale differences if possible. Also align vertical references by applying height corrections using benchmarks. Once these coordinate adjustments are done properly, the point cloud will align almost exactly with the drawing. Make it a habit to measure at least one known point before starting work at a new site to check alignment; if any discrepancy exists, redo localization early.

Q2. What should I do if measured point elevations do not match the drawing’s values? A. In many cases, elevation mismatches are due to differences in reference. GNSS-derived heights are ellipsoidal or geoid heights, while design drawing elevations are often orthometric heights above mean sea level. To reconcile them, measure a known benchmark on site using RTK and calculate the difference between that measurement and the drawing elevation. For example, if benchmark A’s elevation on the drawing is 50.000 m (164.042 ft) and the RTK ellipsoid height is 84.300 m (276.575 ft), the difference is 34.300 m (112.533 ft), which is the geoid separation at that point. Subtracting 34.300 m (112.533 ft) from the Z coordinates of other point cloud points will align them with the drawing’s elevation reference. Modern positioning devices and software often include national geoid models to automatically perform this conversion, which is convenient.

Q3. Can DXF drawings be automatically generated from point cloud data? A. Fully automatic generation of DXF drawings from point clouds is currently difficult. Research into automated point cloud processing is progressing, and there are attempts to detect planar elements like walls and floors for vectorization, but variability across buildings and terrains makes eliminating human judgment entirely impractical. In practice, a realistic approach is for an operator to efficiently trace vector lines over a point cloud used as a background guide. That said, point clouds greatly reduce manual effort compared to measuring by hand. Tools like the LRTK cloud that allow easy drawing on point clouds are available, so using semi-automated tools and finishing manually is the mainstream approach.

Q4. Which format is better to deliver, DXF or DWG? A. If the recipient does not specify or you do not know their software, DXF is a safe choice. DXF is a widely supported format readable by many CAD and GIS programs, offering compatibility. If the recipient uses a specific CAD package and you know the version, you can provide DWG for more complete data retention, but DWG can be unreadable if versions mismatch. When in doubt, provide DXF, or confirm the preferred format with the recipient when DWG is required.

Q5. Do I need expensive equipment to acquire point cloud data? A. Not necessarily. Historically, high-precision 3D laser scanners costing hundreds of thousands of dollars were the norm, but today point clouds can be easily obtained with smartphones and drones. For small-scale indoor measurements, a LiDAR-equipped smartphone is often sufficient; for large-area terrain surveys, photogrammetry with an RTK-enabled drone is an option. Choose the method that fits your requirements—expensive equipment is not always necessary. You can also outsource to specialized surveyors or rent equipment. Start small with accessible tools and scale up as needed.

Q6. Is smartphone LiDAR point cloud accuracy sufficient? A. In many cases, point clouds from smartphones are sufficiently accurate for plan drawing creation and rough quantity estimation. LiDAR sensors in recent smartphones typically yield indoor point clouds with errors on the order of several centimeters depending on the environment, which is adequate for renovation drawings and equipment layout planning. However, for tasks that require millimeter-level precision—such as control-point surveying or steel fabrication measurements—a smartphone alone may be insufficient. In those cases, using RTK-GNSS in combination with LRTK to improve positioning accuracy is effective. A smartphone plus RTK can provide centimeter-class accuracy over wide outdoor areas, achieving results comparable to traditional expensive equipment.