Stereo Video to Point Clouds

Getting Started

clone the repository using git clone https://github.com/wuxiaohua1011/stereo_image_to_point_cloud_japan.git
Setup an identical conda environment using conda create --name myenv --file spec-file.txt
Activate the environment using conda activate myenv
install required modules pip install tqdm opencv-python
Download the data video file mkdir data/raw_data && python download_google_drive.py 1DZvYQMRoPbrmJ67q4kNZ1duFWk5oHnV_ ./data/raw_data/both_eye_0304_1.mp4
run python together.py, you should see a windows pop up, press s to draw a bounding box, and then press space or enter to confirm the bounding box, then you should see both the position and the disparity map showing.

Feel free to disable showing disparity map by passing in False on line 102 for calculating disparity map

Sample Program Run Without Selection

Sample Program Run With Selection

Name		Name	Last commit message	Last commit date
Latest commit History 22 Commits
__pycache__		__pycache__
data		data
old_scripts		old_scripts
paper		paper
past_scripts		past_scripts
related_papers		related_papers
test		test
Japan_Final_Report.pdf		Japan_Final_Report.pdf
README.md		README.md
download_google_drive.py		download_google_drive.py
spec-file.txt		spec-file.txt
stereo_match_multiple.py		stereo_match_multiple.py
stereo_match_multiple_my.py		stereo_match_multiple_my.py
stereo_match_multiple_tim_smart.py		stereo_match_multiple_tim_smart.py
stereo_match_new.py		stereo_match_new.py
together.py		together.py
util.py		util.py
video_to_frames.py		video_to_frames.py