You can match and commit any char after one or more characters that need to be removed in order to recover them later using the link:
sed -i 's/\x02\{1,\}\(.\)/\1/g' foo.txt
Here \x02\{1,\}corresponds to 1 or more characters that need to be deleted, \(.\)will match and capture any char to group 1 and \1will restore the captured char result.