Skip to content

Latest commit

 

History

History
57 lines (37 loc) · 1.81 KB

README.md

File metadata and controls

57 lines (37 loc) · 1.81 KB

videoExtractor

simple script and library for extracting images from video file and adding exif data to the extracted images

There are two ways of doing this, either input the individual values as command line arguements or pass an image from the camera that has exif data that will be mirrored in the extracted video images

extractFrame.py is the manual script that requires a video file and cmd args to work

extractFrameAuto.py is the automated script, needs a video file and image to work

extractFrameAuto_Calibrate.py is the automated script that will undistort image as well as add EXIF information from video images

python extractFrame.py -h will display the arguments you need for input

Tested and Verified on Ubuntu 14.04 LTS

Build docker image

docker build -t ua-videoextractor .

push to ECR

eval $(aws ecr get-login --no-include-email --region ap-southeast-2 | sed 's|https://||')
docker tag ua-videoextractor:latest 712356514005.dkr.ecr.ap-southeast-2.amazonaws.com/video-modelling:ua-videoextractor
docker push 712356514005.dkr.ecr.ap-southeast-2.amazonaws.com/video-modelling:ua-videoextractor

Pulling image from Unleash Hub

Ensure AWS secrets are configured on host

eval $(aws ecr get-login --no-include-email --region ap-southeast-2 | sed 's|https://||')
docker pull 712356514005.dkr.ecr.ap-southeast-2.amazonaws.com/video-modelling:ua-videoextractor

Running image

Input data folder: /data Output data folder: /frames

SOURCE_PATH=`readlink -f data`
docker run -it --mount type=bind,src=${SOURCE_PATH},dst=/data  ua-videoextractor -a 227/100 -focal 4.73 -fnumber 2.2 -cb DJI -cm FC220 -file /data/DJI_0955.MP4 -n 30

interactive mode

docker run -it --mount type=bind,src=${SOURCE_PATH},dst=/data --entrypoint "/bin/bash" ua-videoextractor