php抓网站图片

php抓网站图片的程序

<?php

$fp = fopen('http://www.163.com', 'r');

$data = '';

while(!feof($fp)){

    $data .= fread($fp, 8912);

}

fclose($fp);

preg_match_all( "'<\s*img\s.*?src\s*=\s*

                ([\"\'])?

                (?(1) (.*?)\\1 | ([^\s\>]+))

                ' isx", $data, $links);

while(list($key, $val) = each($links[2])){

    if(!empty($val))

        $match[] = $val;

}

while(list($key, $val) = each($links[3])){

    if(!empty($val))

        $match[] = $val;

}

$count = count($match);

for($i = 0; $i < $count; $i++){

    echo $match[$i].'<br>';

}

?>

用网易主页测试，结果如下：

http://www.163.com/images/neteaselogo.gif

http://www.163.com/images/button_1.gif

http://www.163.com/images/button_3.gif

http://www.163.com/images/button_4.gif

http://www.163.com/images/button_5.gif

http://adimg.163.com/homepage/eachnet/0721/ebay.gif

+ imgSource +

http://cimg2.163.com/sports/2006/8/1/200608011203471f45a.jpg

http://cimg2.163.com/ent/2006/8/1/200608011518238760f.jpg

http://cimg2.163.com/ent/2006/8/1/200608011524370676d.jpg

http://cimg2.163.com/stock/2006/8/1/2006080111352263ed3.jpg

http://cimg2.163.com/biz/2006/8/1/200608011147230c5ec.jpg

http://cimg2.163.com/digi/2006/7/31/nuo/6075.jpg

http://cimg2.163.com/auto/2006/8/1/20060801134705d9391.gif

http://cimg2.163.com/health/2006/8/1/20060801090700ba56b.jpg

http://cimg2.163.com/lady/2006/8/1/200608011026029a8be.jpg

http://adimg.163.com/homepage/eachnet/0801/img01.gif

http://adimg.163.com/homepage/eachnet/0801/lycra.gif

http://adimg.163.com/homepage/eachnet/0801/ym.gif

http://adimg.163.com/homepage/eachnet/0801/camera_36x36.gif

http://adimg.163.com/homepage/eachnet/0801/cloth_36x36.gif

http://adimg.163.com/homepage/eachnet/0801/toy_36x361.gif

http://cimg.163.com/stock/d.gif

http://cimg.163.com/stock/d.gif

http://images.163.com/images/163homepage/biaoshi.gif

http://images.163.com/bj110.gif

http://adgeo.163.com/ad_cookies

惭愧，有两个有问题（红色标记）。

打开网易主页的源代码，有如下两句：

<img src=" + imgSource + " width=185 height=120 alt=点击查看 />

<img src='http://adgeo.163.com/ad_cookies'  width="0" height="0">