Building a SLAM system using microsoft Kinect.

This is my Undergraduate SLAM system which is actually based on ORB_SLAM2 but without loop-closing algorithm. For evaluation, I compared the results of my project with ORB_SLAM and PTAM.

I tried my project on TUM dataset.

If you want to run my code, please drag the 'Thirdparty' and 'Vocabulary' profile provided by ORB_SLAM2 to the project dir.

TUM Dataset

1. Download a sequence from http://vision.in.tum.de/data/datasets/rgbd-dataset/download and uncompress it.

2. Associate RGB images and depth images using the python script associate.py. We already provide associations for some of the sequences in Examples/RGB-D/associations/. You can generate your own associations file executing:

python associate.py rgbd_datatest/rgb.txt rgbd_datatest/depth.txt > associations.txt

python associate.py fr1_desk/rgb.txt fr1_desk/depth.txt > associations.txt

python associate.py rgbd_desk/rgb.txt rgbd_desk/depth.txt > associations.txt

python associate.py rgbd_dataset_slam/rgb.txt rgbd_dataset_slam/depth.txt > associations.txt

python associate.py fr1_xyz/rgb.txt fr1_xyz/depth.txt > associations.txt

3. Execute the following command. Change TUMX.yaml to TUM1.yaml,TUM2.yaml or TUM3.yaml for freiburg1, freiburg2 and freiburg3 sequences respectively. Change PATH_TO_SEQUENCE_FOLDERto the uncompressed sequence folder. Change ASSOCIATIONS_FILE to the path to the coresponding associations file.

./main_tum setting/TUM1.yaml Vocabulary/ORBvoc.txt rgbd_datatest associations.txt

./main_tum setting/TUM1.yaml Vocabulary/ORBvoc.txt fr1_desk associations.txt

./main_tum setting/TUM1.yaml Vocabulary/ORBvoc.txt rgbd_desk associations.txt

./main_tum setting/TUM1.yaml Vocabulary/ORBvoc.txt rgbd_dataset_slam associations.txt

./main_tum setting/TUM1.yaml Vocabulary/ORBvoc.txt fr1_xyz associations.txt

#######
If you want to run the code in qtcreator， just use the cmd starting from setting/... without main

fr1/desk2 for desk_test2
fr2/desk

Drawing trajactory
python draw_groundtruth.py