GeoPointClouds is a 3D Visualization Software created to perform geological interpretations on point cloud data. Version 0.0 functionality consists of:
- File Input and Output for Point Cloud Data Types: .txt 1, .ply, .pcd, .pts
- Graphical Visualization of 3d Point Clouds:
- Ability to rotate and perform zoom and cropping operations on data
- Volume Selection for Point Cloud Labeling:
- Polygon Selection
- Projection Selection
- Line drawing functionality
- Graphical User Interface
All functionality is contained within an application that can be run as an executable on MacOS and Windows. To install the software, see installation. This application requires Python version >3.8 to run.
1 User must specify format for .txt files in the File Import Menu
upon launching the application, you will be see an import file menu to load your file in for use. The components of this menu consist of:
- a
select filebutton: this will launch your file explorer to locate the point cloud file. - a
file formatentry box: if you are using a .txt file, this is where you will define the file heading format, such as 'x y z r g b' or 'x y z r g b nx ny nz' - a
donebutton: click to begin processing your point cloud file for the program..png?raw=true)
.png)
Pro Tip: when importing a .txt file and providing a format (ie. a txt file named pointcloud.txt with format 'x y z r g b'), GeoPointClouds will automatically generate a new file for your data with the extension of its format, such as 'pointcloud.xyzrgb.txt'. This file will generally be smaller and optimized for the program, and you can use it in the future without supplying the format since the format is specified in the extension.
after importing a file for use in GeoPointClouds, you will be directed to the main menu. The main menu has the following functionalities:
Select Volume: see Select VolumeEdit/View Labels: see Edit/View LabelsDraw Line: see Draw LineRender/Export: see Render/ExportView Point Cloud: see View Point Cloud- a
?button: this button will direct you to this repository and readme documentation.
A MacOS 11.0.1 display of the main menu should look like the menu below.
Before clicking Select volume, you can assign or create a label for the volume selected as well as give the area selected a name, description, and color in the Edit/View Labels menu. In the example below, "Sandstone" is selected as the rock type.
Upon clicking Select Volume, you will be directed to an open3D VisualizerWithEditing() window and the main menu for the app will disappear. At this point you will be able to select the points to form your volume selection on the cloud, as shown in the image below. Note that the application will select your volume for labelling based on a bounding polygon you draw around the points you would like to select.
Some pertinent commands for selecting your volume include:
shift + left click: Select a point in the point cloud to begin forming your polygon for volume selection.shift + right click: "Undo" your last point selection in your volume selection.qkey oresckey: Close the window when you are finished making your volume selection.
Note that you must choose at least four points to form a polygon that will select a volume effectively. To view your labeled point cloud after selecting a volume, use View Point Cloud from the main screen.
Upon clicking Edit/View labels for the first time in the application, an empty classification menu will pop up as shown below:
To create a new classification, click the New Classification button and a prompt will appear with the fields "Name", "Classification", and "Color". If "Color" is clicked then a color picker menu will appear as well. These two windows are displayed below.
Once a new classification is added, it will appear in the list and can be selected as the current rock type to be classified by selecting its corresponding radio button to the left of the name. In the example below, two classifications were created, named "Sandstone" and "Mudstone", and the current selected classification for volume selection and labelling is "Sandstone".
Upon clicking Draw Line, an open3D VisualizerWithEditing() will appear and you will be able to select the points to form your lines.
Some pertinent commands for selecting your line's points include:
shift + left click: Select a point in the point cloud to begin forming your polygon for volume selection.shift + right click: "Undo" your last point selection in your volume selection.qkey oresckey: Close the window when you are finished making your volume selection.
An example of line select is shown below. The first image is the user selected points before pressing q to quit the visualizer and choose a color. The second image is from clicking View Point Cloud which now displays the the line set created by connecting the points in order of selection. Note that the +/- button changes the point size in display, which is why point sizes are different in the two examples.
Upon clicking Render/Export, you will be prompted to select how you would like to render the files.
The options for file export include:
- Export a photo of the point cloud as a .png: You will be able to position the camera in open3D to get the angle you would like for the photo, and then press
qorescto close the open3D window and save the photo. - Export colored point cloud: the current point cloud will be exported as a .pcd with all color information retained from the user's interactions with the data.
- Export as separate point cloud: the recolored point cloud will be exported, and an individual point cloud will be exported for each class with all original point information. Note that this option creates several files.
Upon clicking View Point Cloud, an open3D Visualizer() window will display on your screen for displaying the points. To exit this window and go back to the main app, press q or esc.
This application requires Python version >3.8 to run, with pip installed on the computer. The following command block will install all required dependencies for GeoPointClouds.
pip install open3d numpy tk
Next, download this github repository into a folder of your choice.
cd ~/my/file/path/example
git clone https://github.com/nchaconbgeo/pointcloudpackage
To launch the app from here, navigate to the src folder in point cloud package. Assuming you are located in the example directory above (~/my/file/path/example), you would use the commands:
cd pointcloudpackage/src
python3 application.py
If you would like a Desktop App to run the app from, you can follow the instructions on this website.
A video tutorial that outlines the instructions from the website above with specific reference to GeoPointClouds is available here: https://www.youtube.com/watch?v=v9F55mZe7hw
Open3D's documentation has some useful information for navigating the open3D window. A simple tutorial on Open3D visualization can be found here. For this project, all visualized windows are either open3D.visualization.Visualizer() objects and open3D.visualization.VisualizerWithEditing() objects. Hence, these are the types of visualizers that we recommend getting accustomed to and familiar with the documentation on.
Documentation for code can be found either within the code, or consolidated in a sphinx html format in _build/html/index.html