一、前言

之前推荐的 讯飞OCR API额度用掉了，于是转向百度OCR，不得不说还是百度大气：调用量：50000次/天；QPS/并发：2qps。下面以PHP为例介绍百度OCR API构建。

二、API构建

新建index.php、access.ini（存放access_token）、expires.ini（存放token过期时间戳）文件，提前申请好你的百度API Key和Secret Key。

php文件内写入

<?php
/*
*@通过curl方式获取指定的图片到本地
*@ 完整的图片地址
*@ 要存储的文件名
*/
function getImg($url = "", $filename = ""){
    //去除URL连接上面可能的引号
    $hander = curl_init();
    $fp = fopen($filename, 'wb');
    curl_setopt($hander, CURLOPT_URL, $url);
    curl_setopt($hander, CURLOPT_FILE, $fp);
    curl_setopt($hander, CURLOPT_HEADER, 0);
    curl_setopt($hander, CURLOPT_FOLLOWLOCATION, 1);
    curl_setopt($hander, CURLOPT_TIMEOUT, 60);
    curl_exec($hander);
    curl_close($hander);
    fclose($fp);
    return true;
}

/**
* 发起http post请求(REST API), 并获取REST请求的结果
* @param string $url
* @param string $param
* @return - http response body if succeeds, else false.
*/
function request_post($url = '', $param = '')
{
    if (empty($url) || empty($param)) {
        return false;
    }

    $postUrl = $url;
    $curlPost = $param;
    // 初始化curl
    $curl = curl_init();
    curl_setopt($curl, CURLOPT_URL, $postUrl);
    curl_setopt($curl, CURLOPT_HEADER, 0);
    // 要求结果为字符串且输出到屏幕上
    curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
    curl_setopt($curl, CURLOPT_SSL_VERIFYPEER, false);
    // post提交方式
    curl_setopt($curl, CURLOPT_POST, 1);
    curl_setopt($curl, CURLOPT_POSTFIELDS, $curlPost);
    // 运行curl
    $data = curl_exec($curl);
    curl_close($curl);

    return $data;
}
class Sample {
    
    public function run() {
        $curl = curl_init();
        curl_setopt_array($curl, array(
            CURLOPT_URL => "https://aip.baidubce.com/oauth/2.0/token?client_id=【你的API Key】&client_secret=【你的Secret Key】&grant_type=client_credentials",
            CURLOPT_TIMEOUT => 30,
            CURLOPT_RETURNTRANSFER => true,
            CURLOPT_CUSTOMREQUEST => 'POST',
            
            
            CURLOPT_HTTPHEADER => array(
                'Content-Type: application/json',
                'Accept: application/json'
            ),

        ));
        $response = curl_exec($curl);
        curl_close($curl);
        return $response;
    }
}
function get_token(){
    $time = time();
    $rtn = (new Sample())->run();
    //echo $rtn;
    $res = json_decode($rtn, true);
    $access_token = $res["access_token"];
    $expires_in = $res["expires_in"];
    $expires_time = $time + $expires_in;
    
    file_put_contents("expires.ini", $expires_time);
    file_put_contents("access.ini", $access_token);
    
    return $access_token;
}

header('Access-Control-Allow-Origin:*');
header('Content-type: application/json');

$pic=isset($_GET['pic'])? $_GET['pic'] :null; 
if(empty($pic)){die("请传入图片网址参数");}

preg_match('/[^.]+$/', $pic, $matches);
$extension = $matches[0];
$array = ["jpg","jpeg","bmp","png"];
if(!in_array($extension, $array)){die("当前图片格式不支持");}

$expires_time = file_get_contents("expires_time.ini");
$now_time = time();

if($now_time < $expires_time){
    $token = file_get_contents("access_token.ini");
}else{
    $token = get_token();
}

$file_time = gmdate('H-i-s', time() + 3600 * 8);   
$jpg = $file_time.'.'.$extension;
getImg($pic,$jpg);

$url = 'https://aip.baidubce.com/rest/2.0/ocr/v1/general_basic?access_token=' . $token;
$img = file_get_contents($jpg);
$img = base64_encode($img);
$bodys = array(
    'image' => $img
);
$str = request_post($url, $bodys);
$res = json_decode($str, true);
$words = "";
$num = $res['words_result_num'];
for ($i=0; $i<$num; $i++)
{
    $w = $res["words_result"][$i]["words"];
    $words .= $w;
    if($i<($num-1)){
         $words = "$words\n";
    }
   
}

if($words) {$code = '200';}
else{$code = '202';}
$json_return = array(
    "code" => $code,
    "src" => $pic,
    "dst" => $words
);

header('Access-Control-Allow-Origin:*');
header('Content-type: application/json');

echo json_encode($json_return, JSON_UNESCAPED_UNICODE);

unlink($jpg);

三、API调用

请求示例：

 https://api.szfx.top/bdocr/general/?pic=https://api.szfx.top/bdocr/demo.jpg

返回数据：

{
    "code": "200",
    "src": "https://api.szfx.top/bdocr/demo.jpg",
    "dst": "关注【技术松鼠】公众号\n微信搜一搜\n获取优质资源，紧跟时代潮流\nQ技术松鼠"
}

四、构建识别网站

这个比较简单，直接看效果：https://tool.szfx.top/bdocr

 

五、参考资料

一、通用文字识别（标准版）:https://cloud.baidu.com/doc/OCR/s/zk3h7xz52

二、OCR文字识别 – 松鼠API：https://api.szfx.top/api/bdocr.html