# SLAM Process

## 1 Add Scan

The Add Scan function is available only after a project has been created or an existing project has been opened.

{% stepper %}
{% step %}

#### Add R200 Scan Station

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FM6Y1DMfB2VRso4WABa4I%2Fadd1.png?alt=media&#x26;token=f3d7845d-4758-4af3-8b9b-73181d570199" alt=""><figcaption></figcaption></figure>

To add scan data, click **Add Scan** to open the import interface, then click **Add Scan Data** and select the dataset collected by the **R200** hardware system. Both single and multiple scan sessions are supported.

* When importing a single scan session, select the directory containing the original acquisition data.&#x20;
* For batch import, place multiple session folders in the same directory and select that directory to load all sessions at once. It is also possible to use **Ctrl + Left Click** to select multiple individual stations (Figure 60–61).
  {% endstep %}

{% step %}

#### Check Scan Status

After data is added, the system performs an integrity check and displays the scan status.&#x20;

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FT5EOQrg3pPtW5CLy2fNA%2Fcheck.png?alt=media&#x26;token=f40acd29-852b-4591-81d2-c49635801c62" alt=""><figcaption></figcaption></figure>

* **Succeed**: If the data is complete, the status shows **Succeed.**&#x20;
* **Failed**: If required files are missing, the status shows **Failed**. Hovering the cursor over the failed status will indicate the missing file type.&#x20;
  {% endstep %}

{% step %}

#### Operation

Added scan can be removed using the Delete function if needed.
{% endstep %}

{% step %}

#### **Confirm Data**

Click **OK** to complete the scan import. Additional scans may still be added afterward. If duplicate data is imported, the system will prompt that the dataset has already been added and cannot be imported again (Figure 62–63).
{% endstep %}
{% endstepper %}

## 2 Point Cloud Processing

The **Point Cloud Processing** function provides a full workflow for processing raw data collected by **R200** LiDAR scanners. With a single operation, the software can generate multiple outputs, including colorized point clouds, photo-realistic point clouds, Mesh models, and 3DGS models.

Processing settings include point cloud computation and scene type selection, point cloud optimization, preprocessing, and reconstruction. Users can proceed using default parameters for standard workflows or customize parameters according to specific project requirements.

{% hint style="info" %}
**Note:** Successfully completed subtasks are automatically skipped during reprocessing. To recompute the workflow from the beginning, use the **Reset** function first. Changes to earlier processing steps will affect subsequent steps, and modifying upstream settings will cause the downstream operators to be executed again.
{% endhint %}

### 2.1 Scene Type

After the raw data is added, the algorithm performs point cloud computation based on the selected acquisition scene type to achieve optimal processing results. Supported scene types are described as follows:

* **General Scene**: Open environments with spacious ground areas and façades, such as parking lots, large indoor venues, campuses, or plazas.
* **Narrow Scene**: Long and confined environments, such as corridors less than 1 meter wide, mine tunnels, or small indoor spaces.
* **Dynamic Scene**: Environments with significant movement of people or vehicles, such as traffic tunnels, streets during rush hour, or operating shopping malls.

### 2.2 Control Constraint

The software supports position adjustment based on Ground Control Points (GCPs) or RTK data to generate point clouds with absolute coordinates. When GCP-based optimization is applied, the system can automatically export a point cloud accuracy verification report (Figure 66).

#### **2.2.1 Using GCP**

This function transforms the point cloud map into an absolute coordinate system using control points. Control points are introduced during mapping to perform adjustment calculations, resulting in a globally consistent point cloud dataset. Two methods are supported:

* **Pick GCPs from the point cloud** — suitable when GCPs were not surveyed during acquisition.
* **Read GCPs from raw data** — suitable when GCPs were collected during acquisition. At least 3 points are required, with 4 or more recommended.

{% hint style="info" %}
**Note:** The control point reference coordinate file must be prepared and imported by the user. The file format is a TXT file with columns arranged as: **Point ID, X, Y, Z**, separated by spaces.
{% endhint %}

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FLcW3VMaOVuxbTVzv1xXV%2FEdit%20GCP.png?alt=media&#x26;token=4d0b773c-6018-4df7-8701-d223333b70fe" alt="" width="563"><figcaption></figcaption></figure>

In the Data Management panel, right-click the GCP of the corresponding station and select **Edit** to open the **Edit Control Points** panel. Functions are described below:

1. **Open Coordinate File** — Load a GCP file or a previously saved point-pair file.
2. **Save Coordinate File** — Save the configured point-pair relationships as a file.
3. **Pick Reference Point** — Enables point picking in the point cloud. After selecting a row in the GCP list, pick the corresponding reference point in the viewer; the selected point is displayed as a yellow marker.
4. **Add Point** — Add a new row to the list.
5. **Delete Point** — Delete the selected row.

<mark style="background-color:yellow;">**(1) Workflow for “Pick from Point Cloud”**</mark>&#x20;

{% stepper %}
{% step %}
**Import GCP file**

Import the local GCP file into the **Edit Control Points** panel. The system automatically detects and displays the coordinate system.
{% endstep %}

{% step %}
**Match Reference Points**

Select each GCP row, click **Pick Reference Point**, and select the corresponding location in the point cloud.
{% endstep %}

{% step %}
**Confirm Position**

After all points are picked, click **OK**. The system performs GCP optimization and outputs a point cloud in the absolute coordinate system (consistent with the GCP file).
{% endstep %}
{% endstepper %}

<mark style="background-color:yellow;">**(2) Workflow for “Read from Raw Data”**</mark>

Import the local GCP file. The software automatically matches reference points with control points.\
Check the pairs marked **Convert** and click **OK** to save the task.

#### **2.2.2 Using GNSS**

This function improves accuracy by tightly integrating RTK data with LiDAR SLAM results and converts the point cloud into the selected coordinate system. The following conditions must be met:

* The R200 RTK module must be used during acquisition.
* A CORS account must be logged in via the mobile app beforehand.

The system automatically recommends a projected coordinate system based on RTK data. Users may modify it or define a custom coordinate system.

Supported coordinate transformation methods include:

* Direct Method (offset/constant shift)
* Four-Parameter Transformation
* Seven-Parameter Transformation (parameters can be calculated using **Tool → Parameter Calculation**)

**PPK Processing Support:**

PPK processing is supported. Place the processed file into the pos folder of the original project and name it **PPK.txt**. Then select the PPK mode in the interface to perform computation. This is suitable for areas without CORS network coverage.

The transformation name can use either newly calculated parameters or previously saved transformation parameters.

{% hint style="info" %}
**Note:**

* IE software is recommended for PPK processing.
* File columns must be ordered as: Date, GPSTime, Latitude, Longitude, H-Ell, PDOP, HDOP, VDOP, Q, Z-ECEF
* The recommended time interval is **0.2 s**.
  {% endhint %}

### 2.3 Output Setting

The Point Cloud Optimization module automatically refines the dataset according to the acquisition scene and strategy, removing layering artifacts in the map (enabled by default).&#x20;

#### **2.3.1 Dynamic Filtering**

Dynamic Filtering is designed to eliminate moving objects, floating noise, and other artifacts in the point cloud, resulting in a flatter surface and thinner planar structures for improved data quality.

<div><figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2F8gYnVNsuTHzHBzaXIsYF%2Fbefor-filter.png?alt=media&#x26;token=d7870758-477b-434a-aab2-163970f2e6e7" alt=""><figcaption></figcaption></figure> <figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FC97uVVvBMb7aBtI4VxEE%2Fafter-filter.png?alt=media&#x26;token=4c9c189b-3dbc-46e2-bd9b-b26fd34b489c" alt=""><figcaption></figcaption></figure></div>

#### **2.3.2 Point Cloud Smoothing**

Point Cloud Smoothing reduces noise while preserving geometric features. It minimizes errors caused by excessive point thickness and improves overall accuracy, producing a smoother point cloud that better approximates the original surface.

When this option is enabled, the software automatically performs Dynamic Object Removal by default to ensure optimal results.

#### **2.3.3 Coloring**

This function assigns color information to the point cloud by mapping image textures onto the points, making the dataset more vivid and realistic.

When it is selected, Dynamic Filtering and Point Cloud Smoothing must be executed.

#### **2.3.4 Visual Optimization**

Visual Optimization refines image poses to correct color misalignment and achieve higher-quality colorized point clouds.

This function requires additional processing time and disk space.

### 2.4 Point Cloud Reconstruction

Point Cloud Reconstruction supports the generation of Photo-Level Point Clouds, Mesh Models, and 3DGS Models.

* **Dense Point Cloud**: After colorization, the system performs densification and densified color mapping. Compared with standard colorized point clouds, this provides better visual detail, improved geometric structure, and more faithful texture representation.
* **Mesh Model**: Uses the smoothed, colorized point cloud together with visually optimized image poses as input. A surface reconstruction algorithm converts the point cloud into a textured 3D mesh model.
* **3DGS Model**: Based on the smoothed, colorized point cloud and optimized image poses, the system generates a 3DGS model using the 3D Gaussian Splatting algorithm.

## 3 Reset

**Reset Computation** resets the results generated by Point Cloud Processing. After resetting, the data returns to the state obtained from the initial computation of the original input data, and all optimization and processing effects will be removed.

## 4 Alignment

**Point Cloud Alignment** is used to align two point clouds with overlapping areas. One dataset is defined as the **Reference Point Cloud** and the other as the **Floating Point Cloud**. By adding three or more pairs of corresponding points (tie points) in the operation view—ensuring the points are not collinear and are evenly distributed across the survey area—the software automatically performs coarse registration followed by fine registration. This transforms the Floating Point Cloud into the coordinate system of the Reference Point Cloud, producing an aligned result.

1. In the Data Management Panel, left-click while holding **Ctrl** to select two groups of point clouds from different stations.\
   \&#xNAN;*Figure 72 Point Cloud Alignment Data Selection*
2. Click the **Alignment** button to open the alignment interface.
3. Enter the function interface to operation. Detailed steps:

{% stepper %}
{% step %}

#### Select Reference and Floating Point Cloud

Use the drop-down selection box to define the **Reference Point Cloud** and the **Floating Point Cloud**.

* **Reference Point Cloud**: Its coordinate system remains unchanged during alignment.
* **Floating Point Cloud**: Its coordinates will be transformed to match the Reference Point Cloud.
  {% endstep %}

{% step %}

#### Adjust Point Cloud

Adjust the rendering mode of both datasets using **Point Cloud Adjustment** to make corresponding point selection easier.
{% endstep %}

{% step %}

#### Select Corresponding Points

Select three or more pairs of corresponding points in the view window. The index numbers of each pair must match.

To delete a selected point, right-click the point in the view to open the context menu and remove it.
{% endstep %}

{% step %}

#### Preview

After selecting at least three pairs of corresponding points, click **Align** to execute the process. When completed, click the preview window to view the aligned result.
{% endstep %}

{% step %}

#### Confirm

Click **OK** to confirm and save the alignment result. The Floating Point Cloud will then be permanently aligned to the Reference Point Cloud.
{% endstep %}
{% endstepper %}

## 5 Panorama / Photo

**Panorama / Photo** module provides an interactive display mode linking point clouds with images, including **Split-Screen View** and **Panorama View**. Both modes visualize the relationship between point clouds and photos. Click the corresponding function button to switch between the two display modes.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FNj0oUepKEyRo4EmIb9f0%2Fmenu.png?alt=media&#x26;token=cfc68566-3248-420a-b888-03575eb6e8f4" alt=""><figcaption></figcaption></figure>

* **Split-Screen View** simulates a first-person perspective for browsing point clouds or images, allowing users to inspect the alignment between point cloud positions and photo locations.
* **Panorama View** overlays the point cloud with panoramic images, enabling direct measurements on the panorama with point cloud assistance.

To enter Panorama / Photo mode, first open the **Camera** data within the station and set it to visible. The images will appear in the main data view as 3D spheres. Double-click any sphere in the main view to enter Split-Screen or Panorama mode.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2F21eW6cQSzAwmJSlzZBXW%2Fcamera.png?alt=media&#x26;token=f0dc848e-4038-4a55-aea8-372374cbec93" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2Fb9rK3BMpxiEfuchNx1iO%2Fcamera%20display.png?alt=media&#x26;token=3dd82b59-9eaf-4ce5-91f4-ed9da277494c" alt=""><figcaption></figcaption></figure>

In Split-Screen mode, **Point Cloud Linked** is enabled by default. When switching viewpoints, both panels remain synchronized at the same perspective. If unchecked, the point cloud and image views can be navigated independently.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FCaKbgLHgNlEO6JVRVwXy%2Fsplit%20screen.png?alt=media&#x26;token=aa6d2c02-4592-4a76-8dbd-26523a231eec" alt=""><figcaption></figcaption></figure>

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FJwxfzJAwfSeOKEsgnNYS%2Ffull%20screen.png?alt=media&#x26;token=c0033753-f2e5-4efd-ba98-39926f96a644" alt=""><figcaption></figcaption></figure>

Use **Previous Frame** or **Next Frame** to automatically switch viewpoints. You may also click **Play/Pause** to automatically play or pause frame transitions. Keyboard shortcuts are supported:

* Press **A** to switch to the previous frame
* Press **D** to switch to the next frame
* Press **ESC**, or click the **×** in the upper-right corner of the main view, to exit this mode

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FDNvlvz704ecnGwIOcsOy%2Fquery.png?alt=media&#x26;token=a70b1d0a-53b5-4350-bcd4-526480a9cb37" alt=""><figcaption></figcaption></figure>

**Display Radius** sets a visibility range centered on the current viewpoint (default unit: meters). Point cloud data within this radius is displayed normally, while data outside the range is hidden.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2F40L4eogxQYWLWDzZ9iCI%2Fdisplay%20radius.png?alt=media&#x26;token=fb4aa8e3-3bbd-4afd-9378-05cd36915130" alt=""><figcaption></figcaption></figure>

## 6 Export

Select the export feature to output point cloud results.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2FEUYMaF35d3bL9APPfnP5%2Fexport.png?alt=media&#x26;token=e19e1df2-96e2-491d-8cdd-c610ab5010fa" alt=""><figcaption></figcaption></figure>

{% stepper %}
{% step %}

#### Select Point Cloud

In the Data Management Panel, select one or multiple stations.
{% endstep %}

{% step %}

#### Output Setting

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2F6t5OMNzPA7qMhy4RNfcT%2Fsetting.png?alt=media&#x26;token=c738abb1-83ad-4c61-9814-16c4e49318b2" alt=""><figcaption></figcaption></figure>

Click the **Export** button in the menu bar or toolbar to open the Export Data dialog.

In the export dialog, configure the output path, point cloud format (supported formats include LAS, LAZ, TXT, PLY, PCD, and E57), point cloud density, attribute fields, and point cloud coordinates.

For point clouds with absolute coordinates, an alternative coordinate system may be selected for export. If no absolute coordinates exist, only the default coordinate system is available.
{% endstep %}

{% step %}

#### Confirm

After completing the parameter configuration, click **OK** in the Point Cloud Export dialog to export the point cloud data.

<figure><img src="https://2468521665-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXWiwWKjbM3enZNnTnsmE%2Fuploads%2Fnt0K3Zbq3yAW3vR0AtzD%2Fwaiting.png?alt=media&#x26;token=3463af66-2c9d-4521-9d21-ec70d334acf4" alt=""><figcaption></figcaption></figure>
{% endstep %}
{% endstepper %}