5 Reasons the Point Cloud Viewer's Cross-Section Display Is Misaligned | Check Slice Width, Direction, and Rotation

Table of Contents

• Cause 1: Incorrect slice width setting

• Cause 2: Incorrect selection of section orientation

• Cause 3: Reference point misalignment

• Cause 4: Model rotation error

• Cause 5: Insufficient point density

• Ensuring coordinate consistency with LRTK

• FAQ

When handling point cloud data on surveying and construction sites, the cross-section displayed in a point cloud viewer may not appear as expected. Complaints such as the section being offset, the shape appearing unnatural, or the section not matching drawings are often heard when displaying or extracting 3D point clouds in point cloud viewers. This article organizes the main causes of mismatched cross-section displays in point cloud viewers into five categories, and explains common configuration mistakes and field pitfalls for each cause, along with specific methods for checking and remedies. Finally, it also introduces how to ensure coordinate consistency using the simple surveying technique "LRTK" and summarizes the key points for achieving correct cross-section display.

Cause 1: Incorrect Slice Width Setting

In a point cloud viewer, the point cloud to be displayed as a cross-section is "sliced (section extraction)". A slice is a feature that extracts and displays only the points within a region of a certain thickness (slice width) along a predefined plane. If the slice width is inappropriate, the cross-sectional display may be missing parts or include unnecessary areas, preventing an accurate section from being obtained.

For example, when extracting a road cross-section, if the slice width is set too large, extraneous terrain or structures other than the road may be included in the section, whereas if it is too narrow the point cloud can become sparse and the cross-section shape may become indistinguishable. In particular, when the slice width is very small, a slight misalignment can result in almost no points being extracted within the section. In one study, when extracting a cross-section of a structure with a 0.1 m (0.3 ft) gap, setting the slice width to around 0.1 m (0.3 ft) was reported to result in cases where more than half of the point cloud data was cut off. Thus, if the slice width is set too small relative to the dimensions of design elements and the density of the point cloud, important points can be lost and the cross-section can be missing.

On the other hand, making the slice width too large can include other structures or terrain overlapping the section, making the cross-sectional drawing difficult to read. For example, when crossing a road at an oblique angle, if the width is set too wide, roadside trees and adjacent structures may be drawn on the section, causing the information about the road cross-section to be obscured.

How to Verify and Improve

• Confirm the expected cross-section thickness: Determine the thickness of the cross-section you want to extract (road width, thickness of structures, etc.) from the design drawings or site conditions, and set a slice width that matches it. Set it slightly larger than the cross-section thickness and adjust it incrementally.

• Check the display while adjusting the slice width: Extract the cross-section in the viewer and compare results with different slice widths. If the point cloud is discontinuous, widen the width; if there are too many unnecessary points, narrow it. Fine-tune until an appropriate width is found.

• Stepwise cross-section extraction: It can be effective to first take a thicker slice to get a rough sense of the cross-section, then thin the slice and re-extract only the important parts. This allows you to efficiently review both the overall picture and the details.

• Check for overlapping cross-sections: When displaying cross-sections with a wide width, check for overlapping structures. If overlaps are present, either divide the display area as needed and extract cross-sections individually, or adjust the position of the slice plane.

As above, slice width is an important parameter that determines the "thickness" of the cross-section display. If an appropriate width is not set, the true cross-sectional shape cannot be obtained, which will interfere with surveying and verification of the as-built condition. If a mismatch in the cross-section is caused by the slice width, the basic approach is to first review this setting.

Cause 2: Incorrect selection of the section direction

The result of a section depends greatly on the orientation of the plane used to cut it. In point cloud viewers, it is common to have a feature that lets you determine the orientation of the section plane by specifying any two points or an axis, but if you select the wrong section direction, the section will be cut at an angle that deviates from the design intent. This can lead to mistakes such as intending to extract a road’s transverse cross-section but getting a longitudinal section displayed, or cutting a bridge abutment exactly laterally so that a section along the bridge axis does not appear.

A common mistake is to confuse the perpendicular and parallel directions. For example, when you want a cross-section across a road, you need to set the section in the direction perpendicular to the road centerline, but if you mistakenly set it parallel to the road you will only obtain a longitudinal section. Also, when creating a section line by specifying two reference points, you can accidentally select a different point and end up with a section whose angle is slightly off.

How to check and improve

• Compare with drawings and planned lines: Before setting the cross-section, confirm the cutting direction on the drawings. For roads, clearly identify the station line or reference points for the cross-section location and choose a direction perpendicular to their extension.

• Use the viewer's display angle to verify: Visually inspect the point cloud data as a whole and check whether the plane direction set for the cross-section is cutting correctly. Change the camera viewpoint as needed to confirm the cross-section plane is facing the intended direction.

• Try drawing a temporary line on the cross-section: Temporarily slice with the cross-section plane to get a quick look at the cross-section content and see whether the shape matches what you intended. If something is clearly wrong, the cross-section direction is likely incorrect and should be reset.

• Specify two points accurately: When creating a cross-section by specifying two points, take care not to accidentally pick the wrong points. If possible, input surveyed known points or coordinate values to align the direction precisely.

A mistake in selecting the cross-section direction is extremely critical because it changes the shape of the cross-sectional view itself. If you feel the configured cross-section is “not the cross-section you expected,” first question the orientation of the cross-section plane and review its direction before proceeding to causes 3–5.

Cause 3: Reference point misalignment

Even if the position and orientation of the cross-section are specified correctly, the coordinate system of the entire point cloud is misaligned and the cross-section display will not match the expected position on the drawing. This problem occurs when the reference points or origins do not coincide between the point cloud data and the design data (or other survey data).

For example, if you set a control point used on a survey site with a total station but that coordinate is offset from the origin of the actual point cloud data, extracting sections in the viewer will not place them in the positions shown on the design drawings. Also, when stitching together multiple point cloud files, if the reference positions are inconsistent, they will be displayed displaced from their original positional relationships. In particular, when you forget to specify the coordinate system for data acquired by GPS or RTK, or use different datums (local coordinates and global coordinates), the entire model can unknowingly be shifted by tens of cm (tens of in) to several m (several ft).

How to Check and Improve

• Compare with known control points: If known coordinate points placed on the ground (such as markers) or official survey points are included in the point cloud data, check their coordinate values. If there is an offset from the expected positions, a parallel shift of the entire point cloud is necessary.

• Unify coordinate systems and references: Confirm the coordinate system used in the survey plan (geographic coordinates or a local coordinate system), and select the same coordinate system in the point cloud viewer settings. When importing data into surveying instruments or software, check that the origin and coordinate system have not been shifted.

• Use overlay or registration features: If the point cloud viewer can overlay design data (DXF or CAD data, etc.), overlay the as-built point cloud and the design drawings to verify whether control points and important lines align. If you can visually detect offsets during this process, it becomes easier to perform coordinate corrections.

• Precautions when acquiring data: When acquiring point cloud data, take measures to improve coordinate accuracy, such as re-surveying known points or placing local reference points. Using reference points with large errors tends to cause greater coordinate misalignment in downstream processes.

If the point cloud data's reference points are misaligned, the position of the section line will also be significantly shifted. If you feel that the section extraction position does not match the design drawings, review the coordinate system of the entire point cloud and align the reference points. This reference point alignment is especially important when integrating multiple datasets or comparing against the design drawings.

Cause 4: Model rotation error

If the entire point cloud model is tilted before cutting a cross-section, the cross-sectional plane will also be tilted and the cross-sectional view will not be displayed correctly. Model rotation error occurs when the model's orientation (yaw, pitch, roll) is accidentally shifted during acquisition or loading of point cloud data.

For example, when measuring point clouds with mobile LiDAR or a smartphone, if the device is not kept level or the IMU (inertial measurement unit) calibration is incomplete, the entire model can become slightly tilted. In such cases, the cross-sectional plane will no longer correspond to the correct horizontal and vertical directions, causing the road cross-section to appear higher or lower. Also, when loading into a viewer, the X and Y axes can be swapped due to a coordinate-axis rearrangement error, or the data may be imported rotated counterclockwise.

Verification methods and improvements

• Check alignment against known straight lines/planes: Display parts that should clearly be horizontal or vertical, such as building facades, bridge girders, or the ground, and visually check whether the entire model is tilted. For example, for roads verify that the roadway and sidewalks are straight, and for structures confirm that columns and walls are vertical.

• Use the viewer's level/axis display functions: Many viewers have functions to display the XYZ axes or a horizontal plane marker. Enable these and confirm that the point cloud model is aligned in the correct orientation.

• Rotate and reposition the model: If there is an error, use the viewer's controls to rotate and correct the model. Use pan and rotate tools to level the ground, then re-extract cross-sections. Also, when initially importing, pay attention to the coordinate axis settings and avoid selecting the wrong clockwise/counterclockwise orientation.

• Sensor calibration: Review the calibration status of the point cloud capture equipment. If using a smartphone or an IMU-equipped laser scanner, check that the device is correctly calibrated according to the manufacturer's instructions.

A small rotational misalignment of the model may seem insignificant, but it can greatly affect cross-sections. If a transverse cross-section does not appear level and looks distorted, or a longitudinal section has a slope different from what you expected, rechecking the model's orientation can be the key to resolving the issue.

Cause 5: Insufficient point density

If the point cloud data was not captured densely to begin with, then no matter how correct the cross-section's position or orientation is, there will not be enough "points" to produce a satisfactory cross-section display. Insufficient point density occurs when the scan distance is too great, when enough measurement points were not secured, or due to the performance limits of LiDAR or camera sensors.

In cross-sectional views, point clouds from regions near the cross-section plane are projected to form the shape, but if the points are sparse the cross-sectional shape can become jagged or much of it may appear to be missing. This is especially likely to happen in areas that are far from the measuring instrument—such as building attics, deep inside tunnels, or structures that are far above the ground—where points tend to be left out. Lighting and reflection characteristics can also result in a low density of points on concrete surface irregularities or on areas with low reflectivity.

How to Check and Improve

• Check point cloud density in the viewer: Use the viewer's point display settings to emphasize point size and points near the viewpoint, and visually check the density at the location where you want to take a cross-section. If there are many gaps along the cross-section line, you can judge it as insufficient density.

• Re-scan / additional measurements: Perform additional scans in areas found to have low density. For mobile scanners, slow down walking speed or pass the same spot multiple times to increase points. For drones or navigation-based surveying, lower flight altitude, divide the coverage area, and fly in shorter legs to increase the number of points.

• Match sensor performance: Understand the characteristics of the LiDAR or stereo camera you are using, and design measurement routes that account for fewer points in dark areas or on uniform surfaces. Also consider installing reflectors or markers to increase reflectivity.

• Adjust data processing settings: Point reduction may also be caused by point cloud filtering or thinning settings. Check in your viewer or analysis software whether automatic thinning (Voxelization, etc.) is active, and if unnecessary, disable it to maintain high-density data.

If insufficient point density is the cause, the cross-section will not form a proper shape and you will not be able to obtain a "meaningful cross-section." This problem can be prevented if adequate density is ensured during point cloud acquisition, but if it is noticed afterward, address it with additional measurements or data correction.

Ensuring Coordinate Consistency with LRTK

Up to this point we have looked at five causes of mismatched cross-section displays, and one of the latest technologies that helps solve these problems is LRTK (a simple surveying system using a smartphone + RTK positioning). Because LRTK attaches a GNSS receiver to a smartphone to perform point cloud scanning, latitude, longitude, and altitude absolute coordinates are assigned to every scanned point cloud. As a result, point cloud data collected on site and design data can be handled in the same coordinate system, greatly reducing reference point shifts and alignment mismatches.

Using LRTK, you can obtain point clouds with centimeter-level accuracy and embedded positional information with just a single smartphone, without the need to set up control points as complexly as before. For example, in cross-section surveying, if you walk around each station with LRTK to collect points, then later open that data in a point cloud viewer, you can extract cross-sections that are aligned from the outset with the accurate road station positions. In addition, by using LRTK’s cloud processing, automatic point cloud processing, coloring, and cross-section extraction can be performed rapidly, so post-processing adjustments such as point density and model rotation can also be quickly verified on-site.

Furthermore, surveying with LRTK makes verification using known reference points easy. Because absolute coordinates are embedded in the resulting point cloud, the reference points will align when overlaying the data later with design drawings or other point cloud datasets. By introducing LRTK in this way, the process from point cloud acquisition to coordinate alignment becomes seamless, reducing troubles such as "reference point shifts" and "coordinate system mismatches" in cross-section displays. As a result, the accuracy of cross-section extraction in point cloud viewers increases, and on-site verification work becomes significantly more efficient.

In conclusion, as a countermeasure against misalignment in cross-section displays, not only slice width and orientation but also ensuring coordinate consistency is an important point. In particular, by using RTK-capable smartphone surveying systems such as LRTK, point clouds can be obtained easily and with high accuracy, and adjustments with design data become simpler, making it easier to prevent cross-section display issues in advance.

FAQ

Q1. How wide should I set the slice width? A1. Decide the slice width according to the size of the object or structure whose cross-section you want to take. For example, for a road cross-section, refer to the road width and the thickness of the cover soil layer, and it is a good idea to adjust by trying widths several times the actual size (for example about 1-2 m (3.3-6.6 ft)). If the width is too small, points in the point cloud are likely to be missed, so gradually widen it and check.

Q2. I don't understand how to specify the cross-section direction. A2. Select a direction perpendicular to the reference line on the design drawing (such as the road centerline or foundation axis). In the viewer, either click to specify two points for the section location in advance, or set it by numeric input such as X/Y directions. After setting, confirm the orientation of the section plane on the screen, and, if necessary, reconfigure it as many times as needed to align it accurately.

Q3. How can I verify that the coordinates of a point cloud match those of the design drawings? A3. Verify using the positions of known control points or reference points. If your point cloud viewer can overlay the design drawings or CAD data, check whether the target coordinate points or reference markers align. Also, using point clouds that include latitude and longitude information acquired by GPS, such as LRTK, reduces errors because the surveying coordinate system is unified.

Q4. How can I increase point density? A4. Review the scanning conditions. If the object is too far from the instrument, move closer and rescan; if performing mobile surveying, slow your walking speed; or split the data acquisition area to increase the number of points. Placing targets or reflectors in areas with weak reflections is also effective. Also check the software settings to see whether automatic thinning is being applied, and reduce the thinning amount if necessary.

Q5. What specific benefits does using LRTK offer? A5. LRTK is a surveying tool that combines high-precision GNSS with a smartphone, enabling anyone to easily perform point cloud scanning and coordinate surveying simultaneously. Because the resulting point cloud includes absolute coordinates, positional discrepancies with the design are less likely when extracting cross-sections. Also, since cloud processing simplifies data management, lead times for creating cross-sections and performing quality checks can be greatly reduced. It is gaining attention as a support for disaster surveying and for promoting on-site DX.