您应该转换字符串到字节[] 时指定字符集code>，反之亦然。

 字节[]字节= string.getBytes（UTF-8）;
//饲料字节为Base64
 

和

  //从Base64编码字节
串串=新的字符串（字节，UTF-8）;
 

否则该平台的默认编码将使用这不一定是UTF-8本身。

I'm using org.apache.commons.codec.binary.Base64 do decode string which is utf8. Sometimes I get base64 encoded string which after decode looks like for example ^@k��@@. How can I check if base64 is correct or if decoded utf8 string is valid utf8 string?

To clarify. I'm using

public static String base64Decode(String str) {
    try {
        return new String(base64Decode(str.getBytes(Constants.UTF_8)), Constants.UTF_8);
    } catch (UnsupportedEncodingException e) {
         ...
    }
}

public static byte[] base64Decode(byte[] byteArray) {
    return Base64.decodeBase64(byteArray);
}

You should specify the charset during converting String to byte[] and vice versa.

byte[] bytes = string.getBytes("UTF-8");
// feed bytes to Base64

and

// get bytes from Base64
String string = new String(bytes, "UTF-8");

Otherwise the platform default encoding will be used which is not necessarily UTF-8 per se.