GeneA-SLAM Dataset

This is the official dataset for GeneA-SLAM, a SLAM system enhanced with genetic algorithm-based feature point resampling. The dataset is captured using an ORBBEC Astra RGB-D sensor and designed to be compatible with the TUM RGB-D dataset.

📊 Dataset Content

• RGB-D Images: Synchronized color and depth frames with depth registration.
• ROS Bag File: ROS Melodic-compatible (/camera/rgb/image_raw and /camera/depth/image topics).
• Calibration Parameters: Camera intrinsics and distortion coefficients (see below).
• Reference Ground Truth: Trajectory estimates and 3D point clouds from GeneA-SLAM.

🛠️ Compatibility

• Format: Standardized to TUM RGB-D format via OpenNI2 SDK scripts.
• Tools: Directly usable with TUM dataset evaluation tools.

📷 Camera Calibration

fx = 507.292109502834  
fy = 526.636687152693  
cx = 311.8322920110805  
cy = 226.2697397957387  
k1 = 0.1577126194040356  
k2 = -0.9818469537300319  
p1 = -0.009382215272653267  
p2 = 0.001344533621560567  
k3 = 1.579108437584564  

🚀 How to Use

1. Download:

   • Hugging Face Datasets
   • Google Drive

2. Usage with GeneA-SLAM:
   Follow the official repository for SLAM system setup.

📄 License

Released under the GPL-3.0 license, consistent with GeneA-SLAM and ORB-SLAM2.

📢 Cite Us

If you use this dataset, please cite the GeneA-SLAM paper:

@inproceedings{qing2024geneaslam, 
  title={GeneA-SLAM: Enhancing SLAM with Genetic Algorithm-Based Feature Points Re-sampling}, 
  doi={10.1109/icairc64177.2024.10900093}, 
  booktitle={2024 ICAIRC}, 
  author={Qing et al.}, 
  year={2024}
}

---

Repository: GeneA-SLAM on GitHub
Paper: IEEE Xplore