PointCloudDescriptors

A simple command line program to generate relevant descriptors from point cloud data (.PCD files) before exporting the results into Numpy Files (.NPY files) used with python.

---

Table of Contents

• Installation
• Built With
• Basic Usage
• Features
• Contributing
• References
• License

---

Installation

Important : there is no need for installation.

In order to use it, you must follow these steps:

1. Download the latest release from https://github.com/yyaddaden/PointCloudDescriptors/releases,
2. Extract the binary file,
3. Open the command line prompt,
4. Use it with the instructions below (see Basic Usage).

Built With

This project is built using these technologies:

• Programming Language: C++
• Libraries:
  • PCL (Point Cloud Library) with all the dependicies: https://pointcloudlibrary.github.io/
  • CNPY: https://github.com/rogersce/cnpy
• Integrated Development Environment: Microsoft Visual Studio 2019 Community x64

Basic Usage

In order to use the program, four different parameters must be provided:

1. The descriptors to extract:

• 1: Signature of Histograms of Orientations SHOT [1],
• 2: RGB Signature of Histograms of Orientations SHOT RGB [1],
• 3: Unique Shape Context USC [2].

2. The input .pcd file (point cloud data),
3. The output .npy file (numpy file),
4. The search radius (normals and descriptor parameter).

Below a demonstration:

Optional parameter: The fifth parameter allow to perform downsampling to reduce the number of points.

The apple.pcd file used for the demonstration can be found under the folder resources.

It is provided from the RGB-D Object Dataset: https://rgbd-dataset.cs.washington.edu/dataset/rgbd-dataset_pcd_ascii/

Below the display of apple.pcd file in the program CloudCompare:

Features

This command line program (PointCloudDescriptor) includes several features:

1. Feature generation by selecting one of the provided descriptors (see the list above),
2. Export to a format (Numpy File or .NPY) commonly used in Python,
3. Perform downsampling to reduce the complexity of the point cloud (optional fifth parameter).

Contributing

In order to contribue to this projet, there are two options :

• Option 1 : 🍴 Fork this repo!
• Option 2 : 👯 Clone this repo to your local machine using https://github.com/yyaddaden/PointCloudDescriptors.git

References

[1] Salti, S., Tombari, F., & Di Stefano, L. (2014). SHOT: Unique signatures of histograms for surface and texture description. Computer Vision and Image Understanding, 125, 251-264.

[2] Tombari, F., Salti, S., & Di Stefano, L. (2010, October). Unique shape context for 3D data description. In Proceedings of the ACM workshop on 3D object retrieval (pp. 57-62).

License