SuperMap GIS point cloud data processing Q&A

After obtaining the spatial coordinates of each sampling point on the object surface, what is obtained is a collection of points, which is called a "Point Cloud". The point cloud usually obtained by using a 3D laser scanner or a photographic scanner has a relatively large number of points and is relatively dense, which is called a dense point cloud.

2. Point cloud data format

Commonly used storage formats include las, laz, txt, xyz, ply, etc. Point clouds are discrete three-dimensional points. They can not only be used to produce DEM, DOM, and three-dimensional models, but can also be loaded into three-dimensional scenes to display the original appearance of land objects (surface , vegetation, buildings, power lines, poles, substations, etc.).

3. Point cloud data processing

1.iDesktop loads point cloud

iDesktop has the following two methods to load point cloud data.

A way to import point clouds as datasets.

This method only supports point cloud data in txt format. The imported data is used as a three-dimensional point data set. At this time, the digital elevation data DEM can be generated by interpolation.

Generate cache method.

In addition to point cloud data in txt format, point cloud data needs to be generated and cached in this way, and a three-dimensional slice cache layer must be added to the scene to load point cloud data into the scene for browsing.

1) Import point cloud

The data points acquired by lidar are very dense and are also called point cloud data. Lidar data stores information such as X and Y coordinate positions, elevation values, etc. SuperMap iDesktop supports importing radar data in *.txt format into a two-dimensional or three-dimensional data set.

There are three ways to import point cloud data as datasets:

(1) [Start] tab -> [ Data Processing] -> [ Data Import] .

The specific steps are as follows:

In the [ Data Import ] dialog box, in the toolbar, click the [ Add File ] button to add the *.txt format Lidar file to be imported.
After the addition is completed, double-click the cell corresponding to [ File Type ] in the file list and select the file type as [ Radar File ].
Parameter descriptions of the target data source, result dataset, encoding type, import mode, and source file information in the result settings.
Convert parameters to create image pyramid: Check this parameter to create an image pyramid for the data when importing the data.
After the settings are completed, click the [ Import ] button to import the *.txt file of Lidar data.

(2) Select the [ Data Source] to be imported in [ Workspace Manager ] -> right-click the mouse -> [ Import Data Set] .

(3) [Toolbox] -> [ Data Import] -> [ LIDAR] -> [ Import.TXT] . (iDesktopX)

2) Point cloud cache

In order to improve the loading performance of point cloud data in 3D scenes, the point cloud data must be reorganized and a 3D cache generated. In the [ 3D Data] - [Point Cloud] - [Generate Cache ] function in the SuperMap desktop software, not only It has the function of generating three-dimensional cache from point cloud, and also provides the ability to add new point cloud to the existing cache. At the same time, it can also add grouping information to the point cloud and manage it from the data level.

The details of each process of generating 3D cache from point cloud are introduced below. The process is as follows:

Append file path: Check this option to generate additional files based on existing point cloud cache files. If checked, the cache path and cache name cannot be set and will be the path and name of the existing file.
Cache path: Click the button to the right of "Cache Path", select the save path of the cache file in the pop-up "Browse for Folder" dialog box, and click the "OK" button; you can also enter it directly in the text box to save the cache file. path.
Cache Name: Enter a string as the name of the cache.
Pyramid subdivision type: used to set the subdivision type of cache tiles to create tree pyramids. The number of point clouds is huge, and it is necessary to establish a reasonable data organization through spatial indexing to manage point cloud data. The program provides two methods: quadtree and octree.

(1) Quadtree: A two-dimensional spatial index. The root node is divided into four sub-regions according to the spatial range to form four sub-nodes. Quadtree is suitable for spatial ranges with large plane ranges and small height differences that do not require height-level subdivision, such as city-level buildings.

(2) Octree: Octree is a natural extension of quadtree in three-dimensional space. Based on a point in the space, the space is divided into eight regions to form eight subtrees. Octree is suitable for spatial ranges with large height values and small plane range, such as a tall building.

Characteristic value: an attribute field used for layered coloring in the later stage. There are three characteristic values: height, intensity and category. Click the button on the right and select the characteristic value contained in the configuration file in the drop-down menu.
Geometric compression method: The system provides two compression methods, non-compression and regular compression. Non-compression means using the default geometry method.
S3M version: supports two S3M version numbers, S3M1.0 and S3M2.0.
Generate category: Set the category ID that needs to be generated for cache. You can enter multiple category IDs according to your needs. Separate the IDs with semicolons ";". Both Chinese and English semicolons are acceptable.
Tile subdivision type: Set the subdivision method for generating tiles. The default is global subdivision.
Set level: Check this option to set the level. Enter the start layer and end layer respectively. The start layer must not be larger than the end layer.
After setting the above parameters, click the "OK" button to perform the point cloud generation cache operation. If S3MB is selected as the file type, after the generation is completed, a folder named with the cache name will be generated under the specified cache path. The folder contains a .SCP index file and multiple folders containing .s3mb model files. The result is shown below:

3) Load point cloud

After generating the 3D cache, load and browse point cloud data in the following ways:

After creating or opening a scene window, click the [ Ribbon ] > [ Scene ] tab > the button part of the [ Cache ] item in the [ Data ] group; or in the [ Layer Manager ] window, click [ Normal Layer ] , right-click [ Add 3D Tile Cache ].
The [ Open 3D cache file ] dialog box pops up. The [ File type ] under the dialog box defaults to [ All supported 3D caches ], that is, all 3D cache files supported by SuperMap. In addition, users can also click the drop-down button at the back to select the corresponding cache type.
In the [ Open 3D cache file ] dialog box, select the 3D cache file to be loaded, and then click the "Open" button in the dialog box to complete loading of the 3D cache data.
The newly loaded 3D cache data is displayed on the simulated earth in the scene as a 3D layer in the scene. For specific operation instructions, you can search for [ Load 3D Cache ] in the help document. The link to the online help document is as follows: SuperMap iDesktop 10i (2020) Online Help

4) Layered coloring

Right-click on the point cloud cache layer in [ Scene ] and select [ Layer Properties ] - [ Layered Coloring and Flooding Analysis ] to perform layered coloring on point cloud data without color attributes. Its color characteristic value is The characteristic value parameter selected during caching. For example, in the above figure, the characteristic value is height.

2. WebGL loads point cloud

1) Publishing services

Reference article address: SuperMap GIS release 3D service process

2) WebGL loads point cloud

The code for loading the scene in scene.open mode is as follows:

var promise = scene.open(URL_CONFIG.SCENE_POINTCLOUD);

Please refer to the official website for examples:

http://support.supermap.com.cn:8090/webgl/examples/webgl/editor.html#S3MTiles_pointCloud

4. Possible errors and solutions

Question 1 : When the point cloud is generated and cached, the desktop is completely white, but black edges appear in iserver preview and webgl loading. How to remove them?

Problem analysis: Eye Dome lighting (EDL) to highlight the edges of point clouds and enhance depth, its default value is true.
Solution: When loading data in the front-end WebGL, add layer.pointCloudShading.eyeDomeLighting=false and turn off eyeDomeLighting.

Question 2: The black, white and gray colors in the las point cloud software are all white when imported to the desktop.

Problem analysis: color attributes are not stored in alas point cloud data.
Solution: The point cloud tool converts the las format to the txt format and checks whether it has color attributes. The data has no color attributes. If you need to create color effects, you can set the color through layered coloring in iDesktop, or set layering in the front-end WebGL. Color.

Question 3: After the point cloud generation cache is generated, it will display normally only if it is zoomed to the nearest point. When zoomed further, it will be very sparse. When the lod is adjusted to the maximum, it still feels sparse.

Problem analysis: The level is set when generating the cache, which affects the display.
Solution: Do not set [ Start Layer ] and [ End Layer ] when generating point cloud cache.

Question 4: Cannot see the spherical scene added to the point cloud data generation cache

Problem analysis: The coordinate system set by the point cloud does not match the source data.
Solution: The coordinate system setting is incorrect. For example, the provided insertion point unit is meters, and the projected coordinate system should be set with units in meters.

Question 5: The front-end setting of point cloud layered coloring has no effect

Problem analysis: The height range set by the front end does not include the data height range.
Solution: The maximum visible height and minimum visible height of the layered coloring class must include the height range of the data, otherwise you will not see the effect.