Test whether the text is UTF-8 encoding by PHP.

utf8_check.php

<?php
// 
// Test whether the text is UTF-8 encoding
// 
// Result：
//   1 - Has BOM head
//   2 - Pure UTF-8 context
//   3 - More likely to be UTF-8 content
//   4 - Less likely to be UTF-8 content
// 
function utf8_check($text)
{
	$utf8_bom = chr(0xEF).chr(0xBB).chr(0xBF);
	
	// check BOM head
	if (strstr($text, $utf8_bom) === 0)
		return 1;
	
	$text_len = strlen($text);
	
	// UTF-8 format：
	// < 0x80 | 0xxxxxxx
	// < 0xE0 | 110xxxxx 10xxxxxx
	// < 0xF0 | 1110xxxx 10xxxxxx 10xxxxxx
	// < 0xF8 | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
	// < 0xFC | 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
	// < 0xFE | 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
	
	$bad  = 0;
	$good = 0;
	
	$need_check = 0;
	$have_check = 0;
	
	for ($i = 0; $i < $text_len; $i ++) {
		$c = ord($text[$i]);

		if ($need_check > 0) {
			$c = ord($text[$i]);
			$c = ($c >> 6) << 6;
			
			$have_check ++;
			
			// 10xxxxxx ~ 10111111
			if ($c != 0x80) {
				$i -= $have_check;
				$need_check = 0;
				$have_check = 0;
				$bad ++;
			}
			else if ($need_check == $have_check) {
				$need_check = 0;
				$have_check = 0;
				$good ++;
			}
			
			continue;
		}
		
		if ($c < 0x80)      // 0xxxxxxx
			$good ++;
		else if ($c < 0xE0) // 110xxxxx
			$need_check = 1;
		else if ($c < 0xF0) // 1110xxxx
			$need_check = 2;
		else if ($c < 0xF8) // 11110xxx
			$need_check = 3;
		else if ($c < 0xFC) // 111110xx
			$need_check = 4;
		else if ($c < 0xFE) // 1111110x
			$need_check = 5;
		else
			$bad ++;
	}
	
	if ($bad == 0)
		return 2;
	else if ($good > $bad)
		return 3;
	else
		return 4;
}
?>