Dense Accurate Map Building using Neural Networks
git clone https://github.com/mech0ctopus/damNN-vslam.git
cd damNN-vslam
conda create --name damnn --file requirements.txt
conda activate damnn
Download Pre-processed RGB and Depth Images (Re-sized and colorized) Training Images (5.5GB)
Download KITTI Odometry Ground Truth (4MB)
Note: Raw image data is from the KITTI Raw Dataset (synced and rectified) and the KITTI Depth Prediction Dataset (annotated depth maps).
python damnn-vslam.py
cd damNN-vslam
tensorboard --logdir logs