Selenium+Tesseract-OCR智能识别验证码爬取网页数据

1.项目需求描述

通过订单号获取某系统内订单的详细数据，不需要账号密码的登录验证，但有图片验证码的动态识别，将获取到的数据存到数据库。

2.整体思路

  1.通过Selenium技术，无窗口模式打开浏览器

  2.在输入框中动态输入订单号

  3.将图片验证码截图保存到本地

  4.通过Tesseract-OCR技术去本地识别验证码转化为文字

  5.将获取的验证码输入输入框

  6.点击查询获取列表数据

 3.功能实现

1.下载并安装Google浏览器，安装Google驱动chromedriver.exe，获取安装路径，配置在项目中

2.使用Selenium进行浏览器操作

System.setProperty(浏览器驱动, 浏览器驱动安装位置);
ChromeOptions options = new ChromeOptions();
options.addArguments("--headless");                            // 无窗口模式
options.addArguments("--disable-infobars");                    // 禁言消息条
options.addArguments("--disable-extensions");                  // 禁用插件
options.addArguments("--disable-gpu");                         // 禁用GPU
options.addArguments("--no-sandbox");                          // 禁用沙盒模式
options.addArguments("--disable-dev-shm-usage");
options.addArguments("--hide-scrollbars");                     // 隐藏滚动条

WebDriver driver = new ChromeDriver(options);
driver.get(爬取网站URL);
driver.manage().window().setSize(new Dimension(450, 260));     // 设置游览器打开后调整大小
try {
    // 保存IMG图片到本地
    saveImgToLocal(driver);
    Thread.sleep(2000);
    // OCR智能识别验证码
    String codeByOCR = getCodeByOCR();
    if (codeByOCR != null) {
        try {
            WebElement input1 = driver.findElement(By.id(TEXTBOX1));
            input1.sendKeys(code);
            WebElement input2 = driver.findElement(By.id(TEXTBOX2));
            input2.sendKeys(codeByOCR);
            // 获取table数据
            WebElement addButton = driver.findElement(By.id(SELECT_BUTTON));
            addButton.click();
            List<WebElement> tRCollection = driver.findElement(By.id(TABLE_ID)).findElements(By.tagName("tr"));
            for (int t = 1; t < tRCollection.size(); t++) {
                List<WebElement> tDCollection = tRCollection.get(t).findElements(By.tagName("td"));
                VipLogisticsMinHangDetailVo minHangDetailVo = new VipLogisticsMinHangDetailVo();
                minHangDetailVo.setLogistics_number(code);
                for (int i = 0; i < tDCollection.size(); i++) {
                    String text = tDCollection.get(i).getText();
                    switch (i) {
                        case 0:
                            minHangDetailVo.setTime(text);
                        case 1:
                            minHangDetailVo.setOutlet(text);
                        case 2:
                            minHangDetailVo.setOrganization(text);
                        case 3:
                            minHangDetailVo.setEvent(text);
                        case 4:
                            minHangDetailVo.setDetail(text);
                    }
                }
                list.add(minHangDetailVo);
            }
            log.info("验证码识别成功！");
        } catch (Exception e) {
            if (e.toString().contains("错误提示：验证码错误或已过期！")) {
                log.error("验证码识别错误！" + e.toString());                   
            } else if (e.toString().contains("错误提示：请输入验证码！")) {
                log.error("未输入验证码！：" + e.toString());                       
            } else {
                log.error("其他异常：" + e.toString());
            }
        }
    }
    driver.quit();
} catch (Exception e) {
    e.printStackTrace();
}

3.将图片验证码截图保存到本地（截屏法）

private void saveImgToLocal(WebDriver driver) {
    WebElement element = driver.findElement(By.id(img元素ID));
    //创建全屏截图
    WrapsDriver wrapsDriver = (WrapsDriver) element;
    File screen = ((TakesScreenshot) wrapsDriver.getWrappedDriver()).getScreenshotAs(OutputType.FILE);
    try {
        BufferedImage image = ImageIO.read(screen);
        //创建一个矩形使用上面的高度，和宽度
        Point p = element.getLocation();
        //元素坐标
        BufferedImage img = image.getSubimage(p.getX(), p.getY(), element.getSize().getWidth(), element.getSize().getHeight());
        ImageIO.write(img, "png", screen);

        FileUtils.copyFile(screen, new File(保存本地地址 + "imgname.png"));
    } catch (IOException e) {
        e.printStackTrace();
    }
}

4.将图片验证码保存到本地（鼠标法）

private static void saveImgToLocal1(WebDriver driver) {
    Actions action = new Actions(driver);
    action.contextClick(driver.findElement(By.id(img元素ID))).build().perform();
    try {
        Robot robot = new Robot();
        Thread.sleep(1000);

        robot.keyPress(KeyEvent.VK_DOWN);
        Thread.sleep(1000);

        robot.keyPress(KeyEvent.VK_DOWN);
        Thread.sleep(1000);

        robot.keyPress(KeyEvent.VK_ENTER);
        Thread.sleep(1000);
        //释放向下键，不然在此之前的条目将起作用
        robot.keyRelease(KeyEvent.VK_DOWN);
        Thread.sleep(1000);
        //运行保存
        Runtime.getRuntime().exec(SAVE_IMG_EXE);
        Thread.sleep(10000);
    } catch (Exception e) {
        e.printStackTrace();
    }
}

 1 private static void saveImgToLocal1(WebDriver driver) {
 2     Actions action = new Actions(driver);
 3     action.contextClick(driver.findElement(By.id(img元素ID))).build().perform();
 4     try {
 5         Robot robot = new Robot();
 6         Thread.sleep(1000);
 7 
 8         robot.keyPress(KeyEvent.VK_DOWN);
 9         Thread.sleep(1000);
10 
11         robot.keyPress(KeyEvent.VK_DOWN);
12         Thread.sleep(1000);
13 
14         robot.keyPress(KeyEvent.VK_ENTER);
15         Thread.sleep(1000);
16         //释放向下键，不然在此之前的条目将起作用
17         robot.keyRelease(KeyEvent.VK_DOWN);
18         Thread.sleep(1000);
19         //运行保存
20         Runtime.getRuntime().exec(SAVE_IMG_EXE);
21         Thread.sleep(10000);
22     } catch (Exception e) {
23         e.printStackTrace();
24     }
25 }

5.对本地验证码进行OCR识别

private String getCodeByOCR() {
    String result = null;
    File file = new File(本地图片地址);
    if (!file.exists()) {
        if (systemFalg != 1) {
            file.setWritable(true, false);
        }
        file.mkdirs();
    }
    File imageFile = new File(本地图片地址 + "imgname.png");
    if (imageFile.exists()) {
        ITesseract instance = new Tesseract();
        instance.setDatapath(tessdata存放地址);
        try {
            String doOCR = instance.doOCR(imageFile);
            result = replaceBlank(doOCR);
            log.info("解析的验证码为:{}", result != null ? result : "为空！");
        } catch (Exception e) {
            e.printStackTrace();
            log.error("解析验证码异常！");
        }
    } else {
        log.error("解析验证码的文件不存在！");
    }
    return result;
}

综上，该网页的数据就可以获取了。