make_dataset.sh

#!/bin/bash

function printUsageAndExit {
    echo "Usage: bash $0 input_video_folder [output_image_folder num_frames_per_video]"
    exit -1
}

# Ensure ffmpeg is installed
if ! command -v ffmpeg &> /dev/null
then
    echo "ffmpeg not found. Please install ffmpeg to use this script."
    exit -1
fi
echo "cabird num args: $#"
if [ $# -lt 1 ]; then
    printUsageAndExit
fi

ffmpeg_path=$(dirname `which ffmpeg`)
# Video counter
video_counter=0
if [ $# -gt 0 ]; then
    video_path=$1
    if [ ! -d $video_path ]; then  
        echo "cannot find existing input directory: ${video_path}"
	exit -1
    fi
else 
    printUsageAndExit
fi

if [ $# -gt 1 ]; then
    out_frame_path=$2
else
    out_frame_path=$(dirname $1)
    out_frame_folder=$(basename $1)_images
    out_frame_path=${out_frame_path}/${out_frame_folder}
fi
if [ $# -gt 2 ]; then
    frame_nums_per_video=$3
else
    frame_nums_per_video=8
fi

if [ ! -d $out_frame_path ]; then
    echo "$out_frame_path does not exist, creating it"
    mkdir -p $out_frame_path
    if [ ! -d $out_frame_path ]; then
        echo "cannot create diretory: ${out_frame_path}"
	echo -1
    fi
fi

echo "ffmpeg path: $ffmpeg_path"
echo "input video folder: $video_path"
echo "output image folder: $out_frame_path"
echo "frames per video: $frame_nums_per_video"

cd ${ffmpeg_path}
# Process each video file in the current directory
for video in ${video_path}/*.mp4
do  
    echo "cabird video: ${video}"
    duration=$(ffprobe -v error -select_streams v:0 -show_entries format=duration -of csv=p=0 "$video")  
    interval=$(echo "$duration / $frame_nums_per_video" | bc -l)
    filename=$(basename -- "$video")
    prefix="${filename%.*}"
    #extract frame_nums_per_video frames at the same time interval
    ffmpeg -i "$video" -vf "sws_flags=lanczos+accurate_rnd+full_chroma_int:sws_dither=none:param0=5; fps=1/$interval" -vframes ${frame_nums_per_video} "$out_frame_path/${prefix}_%03d.png"
    echo "Extracted frames for $video"
done