$ ffmpeg -i video.mp4 -vf fps=1 img/output%06d.png Then, install FFMPEG by using PPA: $ sudo add-apt-repository ppa:mc3man/trusty-media $ sudo apt-get update $ sudo apt-get install ffmpeg gstreamer0.10-ffmpegĪfter the installation of FFMPEG, terminal (command prompt) will recognize FFMPEG commands, so we can use it from terminal as below. In order to install FFMPEG for Ubuntu 14.04, read the following instructions.įirstly, you need to install the “libav-tools” package: $ sudo apt-get install libav-tools At the end of extraction, there will be 771 images. Now, we will extract a frame as an image for each second of this video file using our python script that uses FFMPEG tool. The length of this video is 12 minutes and 51 seconds. In this post, a video file (available as an mp4 file here) from official YouTube Channel of The Smurfs downloaded for frame extraction step.

This post tells you how to extract these frames from a given video file using FFMPEG which is an open source audio/video conversion tool. A video file is usually composed of several frames (images). One of them is to record a video which contains objects that will be trained in machine learning or deep learning applications, then extract the frames as an image. There are some common techniques for data augmentation.

Python ffmpeg extract frames from video