☰

删除低分辨率的图像

tag-icon • tag-icon linux command-line images
删除低分辨率的图像

我想知道是否有一种方法可以使用命令行工具（Fedora 19）从目录中递归删除分辨率低于我想要的图像？

例如，我想保留水平至少 3000px、垂直至少 2000px 的图像，并删除任何其他图像。

答案1

这是一个可能的纯粹解决方案，取决于用于告知图像分辨率的bash命令：identifyImageMagick

( find . -type f | xargs identify ) | while read i
do
  fn1=${i%%[*}
  fn=${i%% *}
  tail=${i#* * }
  size=${tail%% *}
  xsize=${size%%x*}
  ysize=${size##*x}
  if [[ ( $xsize -le 3000 ) || ( $ysize -le 2000 ) ]]
  then
    rm "$fn"
  fi
done

不过，我建议先进行备份，然后使用更强大的工具来解析（并仔细检查）输出，identify然后再自动删除文件，例如perl。

答案2

扩展@The Vee 的答案，这里有一个一行程序，可以找到您想要的所有文件，前提是您安装了 ImageMagick、awk、sed、grep、cut、xargs 和 find：

( find . -type f | xargs identify ) | grep -E '^[.][^ ]+? [^ ]+? [0-9]+?x[0-9]+? ' | sed -re 's/^([.][^ ]+?) [^ ]+? ([0-9]+?)x([0-9]+?) .*$/\2 \3 \1/g' | while read resolution; do resx=$(echo "$resolution" | awk '{ print $1 }'); resy=$(echo "$resolution" | awk '{ print $2 }'); if [[ ( $resx -lt 3000 ) || ( $resy -lt 2000 ) ]]; then echo "$(echo "$resolution" | cut -d ' ' -f3-)"; fi; done

解释：

(
  # Recursively print filenames in the current directory
  find . -type f | \
  # Pipe them into xargs which runs "identify" for each file
  xargs identify
 ) | \
# Only keep the files that have been identified  
grep -E '^[.][^ ]+? [^ ]+? [0-9]+?x[0-9]+? ' | \
# Extract filename, width and height using RegEx into the format "<width> <height> <filename>". The filename is at the end to account for spaces in the filename which will be handled by using `cut`
sed -re 's/^([.][^ ]+?) [^ ]+? ([0-9]+?)x([0-9]+?) .*$/\2 \3 \1/g' | \
# For each of these lines extracted, do the following
while read resolution; do
  # Extract the width (first number)
  resx=$(echo "$resolution" | awk '{ print $1 }')
  # Extract the height (second number)
  resy=$(echo "$resolution" | awk '{ print $2 }')
  # Do the resolution check (use lower than, instead of lower equals)
  if [[ ( $resx -lt 3000 ) || ( $resy -lt 2000 ) ]]; then
    # Print the filename
    echo "$(echo "$resolution" | cut -d ' ' -f3-)"
  fi
done

上面的代码只打印文件名，因此如果您想删除它们，只需将其输入到管道中即可xargs：

( find . -type f | xargs identify ) | grep -E '^[.][^ ]+? [^ ]+? [0-9]+?x[0-9]+? ' | sed -re 's/^([.][^ ]+?) [^ ]+? ([0-9]+?)x([0-9]+?) .*$/\2 \3 \1/g' | while read resolution; do resx=$(echo "$resolution" | awk '{ print $1 }'); resy=$(echo "$resolution" | awk '{ print $2 }'); if [[ ( $resx -lt 3000 ) || ( $resy -lt 2000 ) ]]; then echo "$(echo "$resolution" | cut -d ' ' -f3-)"; fi; done | xargs rm

如果您想要询问每个文件是否要删除它，请使用：

( find . -type f | xargs identify ) | grep -E '^[.][^ ]+? [^ ]+? [0-9]+?x[0-9]+? ' | sed -re 's/^([.][^ ]+?) [^ ]+? ([0-9]+?)x([0-9]+?) .*$/\2 \3 \1/g' | while read resolution; do resx=$(echo "$resolution" | awk '{ print $1 }'); resy=$(echo "$resolution" | awk '{ print $2 }'); if [[ ( $resx -lt 3000 ) || ( $resy -lt 2000 ) ]]; then echo "$(echo "$resolution" | cut -d ' ' -f3-)"; fi; done | xargs rm --interactive=always

添加-v还将rm打印被删除的文件。

如果您不想像上面描述的那样通过管道传输，您也可以删除内联：

( find . -type f | xargs identify ) | grep -E '^[.][^ ]+? [^ ]+? [0-9]+?x[0-9]+? ' | sed -re 's/^([.][^ ]+?) [^ ]+? ([0-9]+?)x([0-9]+?) .*$/\2 \3 \1/g' | while read resolution; do resx=$(echo "$resolution" | awk '{ print $1 }'); resy=$(echo "$resolution" | awk '{ print $2 }'); if [[ ( $resx -lt 3000 ) || ( $resy -lt 2000 ) ]]; then filetodelete=$(echo "$resolution" | cut -d ' ' -f3-); echo "$filetodelete $resx $resy"; rm -v "$filetodelete"; fi; done

确保您有备份！不要相信上述命令在所有情况下都能 100% 发挥作用。

相关内容