验证码识别之中国移动话费查询

最近公司一个客户要求查询其手里的一批电话号的余额。

而目前我们公司的移动接口已被收回，所以只能去移动网站抓取了。

由于公司觉得自己开发没必要，钱太少，所以所说就花了5000大洋给外包出去做了。

今天听说已开发完了，刚好我最近又准备辞职正在交接，又没事做。所以就看看是否也去做个试试。

所以就在网上查了一下关于验证码的相关文档

依葫芦画瓢，终于将这个功能搞定。

中国移动网站的验证码很简单。

这也是我第一次做验证码识别。

现将代码记录下来，说不定以后用得到，由于只是练手所以代码写得非常之烂，勿拍

只是练习，所以特征码找得也不是很准哈，不过就现在这样识别率也能达95%。

通过这次发现验证码识别还是比较好玩的，接下来继续研究下，说不定以后能赚点小酒钱呢

<?php
/*** 10086验证码识别* * 注：验证码 周边有边框，且最后一位可能会挨着边线* * */class Valite {// 传说中的特征码var $char = array('A'=>'00000011100000110011111000110011001100011001100110001100110010000111111111000001111111110000000000011000','A '=>'0000001110000011001111100011001100110001100110011000110011001000011111111100000111111111000','B'=>'11111111110001111111111000011000011000011000000110001100000011000111000011100001111111100000011111100000','B '=>'1111111111000111111111100001100001100001100000011000110000001100011100001110000111111110000','C'=>'0011111100000011111111000011100001110001100000011000110000001100011000000110000110000110000','D'=>'00111111000000111111110000111000011100011000000110001100000011000011000011000011111111110001111111111000','D '=>'0011111100000011111111000011100001110001100000011000110000001100001100001100001111111111000','E'=>'0011111100000011111111000011101101110001100110011000110011001100001111100110000011110110000','F'=>'11000000000001111111111000111111111100011000000000001100000000000','G'=>'00111111000000111111110011111000011100111000000110011100000011001011000011001111111111111111111111111110','G '=>'0011111100000011111111001111100001110011100000011001110000001100101100001100111111111111111','H' =>'11111111110001111111111000001000000000001000000000001100000000000110000000000011111111110000111111111000','H '=>'1111111111000111111111100000100000000000100000000000110000000000011000000000001111111111000','I'=>'','J'=>'00000000000010000000000001000000000000111111111111111111111111111',' J'=>'0000000000001000000000000111111111111111111111111111','K' =>'111111111100011111111110000000100000000000111000000000110110000000110001100000110000011000010000000110000000000001000','K '=>'1111111111000111111111100000001000000000001110000000001101100000001100011000001100000110000',' K'=>'11111111110001111111111000000010000000000011100000000011011000000011000110000011000001100001000000011000000000000100000000000000010000000000001000000000000111111111111111111111111111','L'=>'','M'=>'','N'=>'11111111110001111111111000001000000000001000000000001100000000000110000000000011111111110000111111111000','O'=>'','P'=>'11111111111111111111111111011000011000011000000110001100000011000111000011100001111111100000011111100000','P '=>'1111111111111111111111111101100001100001100000011000110000001100011100001110000111111110000','Q'=>'00111111000000111111110000111000011100011000000110001100000011000011000011000011111111111111111111111111','Q '=>'0011111100000011111111000011100001110001100000011000110000001100001100001100001111111111111','R'=>'11111111110001111111111000011000000000011000000000001100000000000','S'=>'011100011000011111000110001100110011000110011001100011000111110000110001110000','T'=>'11000000000001111111110000111111111100011000000110001100000011000','U'=>'11111111100001111111111000000000001100000000000110000000000010000000000010000011111111110001111111111000','U '=>'1111111110000111111111100000000000110000000000011000000000001000000000001000001111111111000','V'=>'110000000000011111000000000011111100000000001111100000000000110000000011111000001111110000011111000000001100000000000','V '=>'11000000000001111100000000001111110000000000111110000000000011000000001111100000111111000001111100000000','W'=>'','X' => '10000000010001110000111000011101111000000011110000000001111000000011110111000011100001110001000000001000','X '=> '1000000001000111000011100001110111100000001111000000000111100000001111011100001110000111000','Y'=>'110000000000011111000000000011111000001000001111111100000000111100000011111000001111100000011111000000001100000000000','Y '=>'11000000000001111100000000001111100000100000111111110000000011110000001111100000111110000001111100000000','Z'=>'1100000111000110000111100011000110110001100110011000110110001100011110000110001110000011000',);/* 读取验证码图片，并将像素RGB读入二维数组，然后分离出有效字符 */function fixImg($url){$img = imagecreatefromjpeg($url);$with = imagesx($img); //取得图像宽度$height = imagesy($img);//取得图像高度$middle = ($height-2) / 2;    //中线$colors = array();/*** 由于有边框，所有从1到宽度-1*/for ($x = 1; $x < $with; ++$x) {for ($y = 1; $y < $height-1; ++$y) {//由于最后一位挨边了，所以最后增加一个空列if( $x == $with-1){$colors[$y][$x] = '0';continue;}$color = imagecolorat($img, $x, $y);$cols = imagecolorsforindex($img, $color);if( $cols['red'] < 125 || $cols['green']<125|| $cols['blue'] < 125 ){$colors[$y][$x] = '1';} else {$colors[$y][$x] = '0';}}}return $colors;}function getHec($imgUrl){$arr = $this->fixImg($imgUrl);$topLine = 6; //上面的空行 $y = 0;// 采用从上到下，从左到右顺序读特征码，从第0行开始（实际是第x行）$len = 19;// 多读一行，JQ超过20行，且J后面几行占了上个字符的位置$code = array();// 分离出来的字符特征码$str = '';$HecCode = '';for($i = 0;$i < $topLine;$i++)unset($arr[$i]);// 前几行是空白while($y < count($arr[$topLine])){$flag = true;// 全0是空白竖线$line = '';for($i = $topLine;$i < $len;$i++){if($arr[$i][$y]){    //第10行，第Y列$flag = false;}$line .= $arr[$i][$y];}if($flag){if(strlen($str) > $len-$topLine)$code[] = $str;$str = '';}else{$str .= $line;}$y++;}$code[] = $str;// 输出字符foreach($code as $v){$match = false;foreach($this->char as $key => $v2){if($v === $v2 && $v!=''){//echo trim($key);$HecCode .= trim($key);$match = true;break;}}}return $HecCode;}}//保存验证码图片
/*** **/
$referer = "http://service.cq.10086.cn/app?service=page/newLogin.login";
$url = "http://service.cq.10086.cn/icsimage?mode=validate&width=51&height=20&temp=";
$cookie = dirname(__FILE__)."/valid_cookie.tmp";$curl = curl_init($url);
curl_setopt($curl, CURLOPT_REFERER, $referer);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($curl, CURLOPT_COOKIEFILE, $cookie);
curl_setopt($curl, CURLOPT_COOKIEJAR, $cookie);
$data = curl_exec($curl);
curl_close($curl);//保存图片到本地
$fp = fopen("valid.jpg","wb");
fwrite($fp, $data);
fclose($fp);/**/
//识别验证码图片
$valid = new Valite();
$code = $valid->getHec('valid.jpg');//要查询的手机号码
$tel = "13888888888";
//手机的服务密码
$pwd = "666666";//提交数据
$postData = "EFFICACY_CODE=".$code."&Form0=blogin&SERIAL_NUMBER=".$tel."&USER_PASSSMS=&USER_PASSWD=".$pwd."&USER_PASSWD_SELECT=1&clogin=".urlencode('on&service=direct/1/newLogin.login/$Form')."&sp=S0";$ch = curl_init();//初始化curl
curl_setopt($ch,CURLOPT_URL,'http://service.cq.10086.cn/app?service=page/newLogin.login&listener=login');//抓取指定网页
curl_setopt($ch, CURLOPT_HEADER, 0);//设置header
curl_setopt($ch, CURLOPT_REFERER, $referer);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);//要求结果为字符串且输出到屏幕上
curl_setopt($ch, CURLOPT_COOKIEFILE, $cookie);
curl_setopt($ch, CURLOPT_POST, 1);//post提交方式
curl_setopt($ch, CURLOPT_POSTFIELDS, $postData);
$data = curl_exec($ch);//运行curl
curl_close($ch);
//print_r($data);//输出结果//抓取结果
$url = "http://service.cq.10086.cn/app?service=page/feequery.QueryBalance&listener=initPage&subSysCode=E003";$curl = curl_init($url);
curl_setopt($curl, CURLOPT_REFERER, $referer);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($curl, CURLOPT_COOKIEFILE, $cookie);
$src = curl_exec($curl);
curl_close($curl);
preg_match('/<td style="text\-align:center;">([0-9\.]+)元<\/td>/', iconv('gbk','utf-8', $src), $res);
print_r("电话号码：".$tel."<br />");
print_r("当前余额：".$res[1].'元');

验证码识别之中国移动话费查询相关推荐

中国移动话费查询，短信查询，各种免费查询！以后别打10086了
发送YE/CXYE到10086,余额查询发送CXBX到10086,查询当月套餐剩余短信条数. 发送CXGFX到10086,查询当月飞信GPRS套餐剩余流量. 发送CXGTC到10086,查询当月GP ...
验证码识别（Python）
图形验证码识别图形验证码最早出现,也很常见,一般由4个字母或者数字组成.例如,中国知网注册页面采用的就是图形验证码,链接为http://my.cnki.net/elibregister/common ...
实战：CNN+BLSTM+CTC的验证码识别从训练到部署 | 技术头条
作者|_Coriander 转载自Jerry的算法和NLP(ID: gh_36eba310d433) 1.前言本项目适用于Python3.6,GPU>=NVIDIA GTX1050Ti,原ma ...
基于Python的验证码识别技术
基于Python的验证码识别技术作者:强哥概述前言准备工作识别原理图像处理切割图像人工标注训练数据检测结果搞笑一刻福利一刻推荐阅读前言很多网站登录都需要输入验证码,如果要 ...
验证码识别，发票编号识别（转）
毕业设计做了一个简单的研究下验证码识别的问题,并没有深入的研究,设计图形图像的东西,水很深,神经网络,机器学习,都很难.这次只是在传统的方式下分析了一次. 今年工作之后再也没有整理过,前几天一个家伙要 ...
[验证码识别技术]字符验证码杀手--CNN
字符验证码杀手--CNN 1 abstract 目前随着深度学习,越来越蓬勃的发展,在图像识别和语音识别中也表现出了强大的生产力.对于普通的深度学习爱好者来说,一上来就去跑那边公开的大型数据库,比如I ...
网络爬虫中的验证码识别
网络爬虫遇到的验证码在写网络,爬虫时,遇到很多网站存在验证码的情形,有其是比较烦的是,爬取数据的每一页都有验证码,如果只有登陆时,存在验证码,这个很好解决,只需将验证码获取后手动输入就行. 但对于每 ...
基于CRNN的文本字符交易验证码识别--Paddle实战
基于CRNN的文本字符验证码识别本项目链接,欢迎大家Fork:https://aistudio.baidu.com/aistudio/projectdetail/3501451 Paddle学习资料 ...
【Matlab验证码识别】遗传算法和最大熵优化+大津法（OTSU）+自定义阈值数字验证码识别【含GUI源码 1694期】
一.代码运行视频(哔哩哔哩) [Matlab验证码识别]遗传算法和最大熵优化+大津法(OTSU)+自定义阈值数字验证码识别[含GUI源码 1694期] 二.matlab版本及参考文献 1 matlab ...

验证码识别之中国移动话费查询

验证码识别之中国移动话费查询相关推荐

最新文章

热门文章