fix: handle multibyte characters in scenario label correctly

[KeitaHiguchi]

Problem

Scenario labels containing multibyte characters (e.g., Japanese, Chinese) were not processed correctly.

For example, when defining a scenario like:

{
  "scenarios": [
    {
      "label": "BackstopJS ホームページ",
      ** other configurations **
    },
    {
      "label": "BackstopJS テストページ",
      ** other configurations **
    }
  ]
}

Multibyte characters are ignored in the generated capture filenames,
causing both scenarios to generate the same filename, making it impossible to differentiate between test results.

Solution

• Fixed the encoding issue when handling scenario labels to ensure multibyte characters are properly retained in capture filenames.