selenium+java破解极验滑动验证码的示例代码

2018-03-12 17:45 钱先生阅读(7036) 评论(2) 编辑收藏举报

转自: https://www.jianshu.com/p/1466f1ba3275

selenium+java破解极验滑动验证码

卧颜沉默关注

2017.08.15 20:07* 字数 3085 阅读 2548评论 2喜欢 10

摘要

分析验证码素材图片混淆原理，并采用selenium模拟人拖动滑块过程，进而破解验证码。

人工验证的过程

打开威锋网注册页面（https://passport.feng.com/?r=user/register）
移动鼠标至小滑块，一张完整的图片会出现（如下图1）
点击鼠标左键，图片中间会出现一个缺块（如下图2）
移动小滑块正上方图案至缺块处
验证通过

selenium模拟验证的过程

加载威锋网注册页面（https://passport.feng.com/?r=user/register）
下载图片1和缺块图片2
根据两张图片的差异计算平移的距离x
模拟鼠标点击事件，点击小滑块向右移动x
验证通过

详细分析

打开chrome浏览器控制台，会发现图1所示的验证码图片并不是极验后台返回的原图。而是由多个div拼接而成（如下图3）

通过图片显示div的style属性可知，极验后台把图片进行切割加错位处理。把素材图片切割成10 * 58大小的52张小图，再进行错位处理。在网页上显示的时候，再通过css的background-position属性对图片进行还原。以上的图1和图2都是经过了这种处理。在这种情况下，使用selenium模拟验证是需要对下载的验证码图片进行还原。如上图3的第一个div.gt_cut_fullbg_slice标签，它的大小为10px * 58px，其中style属性为background-image: url("http://static.geetest.com/pictures/gt/969ffa43c/969ffa43c.webp"); background-position: -157px -58px;会把该属性对应url的图片进行一个平移操作，以左上角为参考，向左平移157px，向上平移58px，图片超出部分不会显示。所以上图1所示图片是由26 * 2个10px * 58px大小的div组成（如下图4）。每一个小方块的大小58 * 10

下载图片并还原，上一步骤分析了图片具体的混淆逻辑，具体还原图片的代码实现如下，主要逻辑是把原图裁剪为52张小图，然后拼接成一张完整的图。

 1 /**
 2  *还原图片
 3  * @param type
 4  */
 5 private static void restoreImage(String type) throws IOException {
 6     //把图片裁剪为2 * 26份
 7     for(int i = 0; i < 52; i++){
 8         cutPic(basePath + type +".jpg"
 9                 ,basePath + "result/" + type + i + ".jpg", -moveArray[i][0], -moveArray[i][1], 10, 58);
10     }
11     //拼接图片
12     String[] b = new String[26];
13     for(int i = 0; i < 26; i++){
14         b[i] = String.format(basePath + "result/" + type + "%d.jpg", i);
15     }
16     mergeImage(b, 1, basePath + "result/" + type + "result1.jpg");
17     //拼接图片
18     String[] c = new String[26];
19     for(int i = 0; i < 26; i++){
20         c[i] = String.format(basePath + "result/" + type + "%d.jpg", i + 26);
21     }
22     mergeImage(c, 1, basePath + "result/" + type + "result2.jpg");
23     mergeImage(new String[]{basePath + "result/" + type + "result1.jpg",
24             basePath + "result/" + type + "result2.jpg"}, 2, basePath + "result/" + type + "result3.jpg");
25     //删除产生的中间图片
26     for(int i = 0; i < 52; i++){
27         new File(basePath + "result/" + type + i + ".jpg").deleteOnExit();
28     }
29     new File(basePath + "result/" + type + "result1.jpg").deleteOnExit();
30     new File(basePath + "result/" + type + "result2.jpg").deleteOnExit();
31 }
32 
33 作者：卧颜沉默
34 链接：https://www.jianshu.com/p/1466f1ba3275
35 來源：简书
36 著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

还原过程需要注意的是，后台返回错位的图片是312 * 116大小的。而网页上图片div的大小是260 * 116。

计算平移距离，遍历图片的每一个像素点，当两张图的R、G、B之差的和大于255，说明该点的差异过大，很有可能就是需要平移到该位置的那个点，代码如下。

 1 BufferedImage fullBI = ImageIO.read(new File(basePath + "result/" + FULL_IMAGE_NAME + "result3.jpg"));
 2     BufferedImage bgBI = ImageIO.read(new File(basePath + "result/" + BG_IMAGE_NAME + "result3.jpg"));
 3     for (int i = 0; i < bgBI.getWidth(); i++){
 4         for (int j = 0; j < bgBI.getHeight(); j++) {
 5             int[] fullRgb = new int[3];
 6             fullRgb[0] = (fullBI.getRGB(i, j)  & 0xff0000) >> 16;
 7             fullRgb[1] = (fullBI.getRGB(i, j)  & 0xff00) >> 8;
 8             fullRgb[2] = (fullBI.getRGB(i, j)  & 0xff);
 9 
10             int[] bgRgb = new int[3];
11             bgRgb[0] = (bgBI.getRGB(i, j)  & 0xff0000) >> 16;
12             bgRgb[1] = (bgBI.getRGB(i, j)  & 0xff00) >> 8;
13             bgRgb[2] = (bgBI.getRGB(i, j)  & 0xff);
14             if(difference(fullRgb, bgRgb) > 255){
15                 return i;
16             }
17         }
18     }
19 
20 作者：卧颜沉默
21 链接：https://www.jianshu.com/p/1466f1ba3275
22 來源：简书
23 著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

模拟鼠标移动事件，这一步骤是最关键的步骤，极验验证码后台正是通过移动滑块的轨迹来判断是否为机器所为。整个移动轨迹的过程越随机越好，我这里提供一种成功率较高的移动算法，代码如下。

 1 public static void move(WebDriver driver, WebElement element, int distance) throws InterruptedException {
 2         int xDis = distance + 11;
 3         System.out.println("应平移距离：" + xDis);
 4         int moveX = new Random().nextInt(8) - 5;
 5         int moveY = 1;
 6         Actions actions = new Actions(driver);
 7         new Actions(driver).clickAndHold(element).perform();
 8         Thread.sleep(200);
 9         printLocation(element);
10         actions.moveToElement(element, moveX, moveY).perform();
11         System.out.println(moveX + "--" + moveY);
12         printLocation(element);
13         for (int i = 0; i < 22; i++){
14             int s = 10;
15             if (i % 2 == 0){
16                 s = -10;
17             }
18             actions.moveToElement(element, s, 1).perform();
19             printLocation(element);
20             Thread.sleep(new Random().nextInt(100) + 150);
21         }
22 
23         System.out.println(xDis + "--" + 1);
24         actions.moveByOffset(xDis, 1).perform();
25         printLocation(element);
26         Thread.sleep(200);
27         actions.release(element).perform();
28     }
29 
30 作者：卧颜沉默
31 链接：https://www.jianshu.com/p/1466f1ba3275
32 來源：简书
33 著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

完整代码如下

  1 package com.github.wycm;
  2 
  3 import org.apache.commons.io.FileUtils;
  4 import org.jsoup.Jsoup;
  5 import org.jsoup.nodes.Document;
  6 import org.jsoup.nodes.Element;
  7 import org.jsoup.select.Elements;
  8 import org.openqa.selenium.By;
  9 import org.openqa.selenium.Point;
 10 import org.openqa.selenium.WebDriver;
 11 import org.openqa.selenium.WebElement;
 12 import org.openqa.selenium.chrome.ChromeDriver;
 13 import org.openqa.selenium.interactions.Actions;
 14 import org.openqa.selenium.support.ui.ExpectedCondition;
 15 import org.openqa.selenium.support.ui.WebDriverWait;
 16 
 17 import javax.imageio.ImageIO;
 18 import javax.imageio.ImageReadParam;
 19 import javax.imageio.ImageReader;
 20 import javax.imageio.stream.ImageInputStream;
 21 import java.awt.*;
 22 import java.awt.image.BufferedImage;
 23 import java.io.File;
 24 import java.io.FileInputStream;
 25 import java.io.IOException;
 26 import java.net.URL;
 27 import java.util.Iterator;
 28 import java.util.Random;
 29 import java.util.regex.Matcher;
 30 import java.util.regex.Pattern;
 31 
 32 public class GeettestCrawler {
 33     private static String basePath = "src/main/resources/";
 34     private static String FULL_IMAGE_NAME = "full-image";
 35     private static String BG_IMAGE_NAME = "bg-image";
 36     private static int[][] moveArray = new int[52][2];
 37     private static boolean moveArrayInit = false;
 38     private static String INDEX_URL = "https://passport.feng.com/?r=user/register";
 39     private static WebDriver driver;
 40 
 41     static {
 42         System.setProperty("webdriver.chrome.driver", "D:/dev/selenium/chromedriver_V2.30/chromedriver_win32/chromedriver.exe");
 43         if (!System.getProperty("os.name").toLowerCase().contains("windows")){
 44             System.setProperty("webdriver.chrome.driver", "/Users/wangyang/workspace/selenium/chromedriver_V2.30/chromedriver");
 45         }
 46         driver = new ChromeDriver();
 47     }
 48 
 49     public static void main(String[] args) throws InterruptedException {
 50         for (int i = 0; i < 10; i++){
 51             try {
 52                 invoke();
 53             } catch (IOException e) {
 54                 e.printStackTrace();
 55             } catch (InterruptedException e) {
 56                 e.printStackTrace();
 57             }
 58         }
 59         driver.quit();
 60     }
 61     private static void invoke() throws IOException, InterruptedException {
 62         //设置input参数
 63         driver.get(INDEX_URL);
 64 
 65         //通过[class=gt_slider_knob gt_show]
 66         By moveBtn = By.cssSelector(".gt_slider_knob.gt_show");
 67         waitForLoad(driver, moveBtn);
 68         WebElement moveElemet = driver.findElement(moveBtn);
 69         int i = 0;
 70         while (i++ < 15){
 71             int distance = getMoveDistance(driver);
 72             move(driver, moveElemet, distance - 6);
 73             By gtTypeBy = By.cssSelector(".gt_info_type");
 74             By gtInfoBy = By.cssSelector(".gt_info_content");
 75             waitForLoad(driver, gtTypeBy);
 76             waitForLoad(driver, gtInfoBy);
 77             String gtType = driver.findElement(gtTypeBy).getText();
 78             String gtInfo = driver.findElement(gtInfoBy).getText();
 79             System.out.println(gtType + "---" + gtInfo);
 80             /**
 81              * 再来一次：
 82              * 验证失败：
 83              */
 84             if(!gtType.equals("再来一次:") && !gtType.equals("验证失败:")){
 85                 Thread.sleep(4000);
 86                 System.out.println(driver);
 87                 break;
 88             }
 89             Thread.sleep(4000);
 90         }
 91     }
 92 
 93     /**
 94      * 移动
 95      * @param driver
 96      * @param element
 97      * @param distance
 98      * @throws InterruptedException
 99      */
100     public static void move(WebDriver driver, WebElement element, int distance) throws InterruptedException {
101         int xDis = distance + 11;
102         System.out.println("应平移距离：" + xDis);
103         int moveX = new Random().nextInt(8) - 5;
104         int moveY = 1;
105         Actions actions = new Actions(driver);
106         new Actions(driver).clickAndHold(element).perform();
107         Thread.sleep(200);
108         printLocation(element);
109         actions.moveToElement(element, moveX, moveY).perform();
110         System.out.println(moveX + "--" + moveY);
111         printLocation(element);
112         for (int i = 0; i < 22; i++){
113             int s = 10;
114             if (i % 2 == 0){
115                 s = -10;
116             }
117             actions.moveToElement(element, s, 1).perform();
118 //            printLocation(element);
119             Thread.sleep(new Random().nextInt(100) + 150);
120         }
121 
122         System.out.println(xDis + "--" + 1);
123         actions.moveByOffset(xDis, 1).perform();
124         printLocation(element);
125         Thread.sleep(200);
126         actions.release(element).perform();
127     }
128     private static void printLocation(WebElement element){
129         Point point  = element.getLocation();
130         System.out.println(point.toString());
131     }
132     /**
133      * 等待元素加载，10s超时
134      * @param driver
135      * @param by
136      */
137     public static void waitForLoad(final WebDriver driver, final By by){
138         new WebDriverWait(driver, 10).until(new ExpectedCondition<Boolean>() {
139             public Boolean apply(WebDriver d) {
140                 WebElement element = driver.findElement(by);
141                 if (element != null){
142                     return true;
143                 }
144                 return false;
145             }
146         });
147     }
148 
149     /**
150      * 计算需要平移的距离
151      * @param driver
152      * @return
153      * @throws IOException
154      */
155     public static int getMoveDistance(WebDriver driver) throws IOException {
156         String pageSource = driver.getPageSource();
157         String fullImageUrl = getFullImageUrl(pageSource);
158         FileUtils.copyURLToFile(new URL(fullImageUrl), new File(basePath + FULL_IMAGE_NAME + ".jpg"));
159         String getBgImageUrl = getBgImageUrl(pageSource);
160         FileUtils.copyURLToFile(new URL(getBgImageUrl), new File(basePath + BG_IMAGE_NAME + ".jpg"));
161         initMoveArray(driver);
162         restoreImage(FULL_IMAGE_NAME);
163         restoreImage(BG_IMAGE_NAME);
164         BufferedImage fullBI = ImageIO.read(new File(basePath + "result/" + FULL_IMAGE_NAME + "result3.jpg"));
165         BufferedImage bgBI = ImageIO.read(new File(basePath + "result/" + BG_IMAGE_NAME + "result3.jpg"));
166         for (int i = 0; i < bgBI.getWidth(); i++){
167             for (int j = 0; j < bgBI.getHeight(); j++) {
168                 int[] fullRgb = new int[3];
169                 fullRgb[0] = (fullBI.getRGB(i, j)  & 0xff0000) >> 16;
170                 fullRgb[1] = (fullBI.getRGB(i, j)  & 0xff00) >> 8;
171                 fullRgb[2] = (fullBI.getRGB(i, j)  & 0xff);
172 
173                 int[] bgRgb = new int[3];
174                 bgRgb[0] = (bgBI.getRGB(i, j)  & 0xff0000) >> 16;
175                 bgRgb[1] = (bgBI.getRGB(i, j)  & 0xff00) >> 8;
176                 bgRgb[2] = (bgBI.getRGB(i, j)  & 0xff);
177                 if(difference(fullRgb, bgRgb) > 255){
178                     return i;
179                 }
180             }
181         }
182         throw new RuntimeException("未找到需要平移的位置");
183     }
184     private static int difference(int[] a, int[] b){
185         return Math.abs(a[0] - b[0]) + Math.abs(a[1] - b[1]) + Math.abs(a[2] - b[2]);
186     }
187     /**
188      * 获取move数组
189      * @param driver
190      */
191     private static void initMoveArray(WebDriver driver){
192         if (moveArrayInit){
193             return;
194         }
195         Document document = Jsoup.parse(driver.getPageSource());
196         Elements elements = document.select("[class=gt_cut_bg gt_show]").first().children();
197         int i = 0;
198         for(Element element : elements){
199             Pattern pattern = Pattern.compile(".*background-position: (.*?)px (.*?)px.*");
200             Matcher matcher = pattern.matcher(element.toString());
201             if (matcher.find()){
202                 String width = matcher.group(1);
203                 String height = matcher.group(2);
204                 moveArray[i][0] = Integer.parseInt(width);
205                 moveArray[i++][1] = Integer.parseInt(height);
206             } else {
207                 throw new RuntimeException("解析异常");
208             }
209         }
210         moveArrayInit = true;
211     }
212     /**
213      *还原图片
214      * @param type
215      */
216     private static void restoreImage(String type) throws IOException {
217         //把图片裁剪为2 * 26份
218         for(int i = 0; i < 52; i++){
219             cutPic(basePath + type +".jpg"
220                     ,basePath + "result/" + type + i + ".jpg", -moveArray[i][0], -moveArray[i][1], 10, 58);
221         }
222         //拼接图片
223         String[] b = new String[26];
224         for(int i = 0; i < 26; i++){
225             b[i] = String.format(basePath + "result/" + type + "%d.jpg", i);
226         }
227         mergeImage(b, 1, basePath + "result/" + type + "result1.jpg");
228         //拼接图片
229         String[] c = new String[26];
230         for(int i = 0; i < 26; i++){
231             c[i] = String.format(basePath + "result/" + type + "%d.jpg", i + 26);
232         }
233         mergeImage(c, 1, basePath + "result/" + type + "result2.jpg");
234         mergeImage(new String[]{basePath + "result/" + type + "result1.jpg",
235                 basePath + "result/" + type + "result2.jpg"}, 2, basePath + "result/" + type + "result3.jpg");
236         //删除产生的中间图片
237         for(int i = 0; i < 52; i++){
238             new File(basePath + "result/" + type + i + ".jpg").deleteOnExit();
239         }
240         new File(basePath + "result/" + type + "result1.jpg").deleteOnExit();
241         new File(basePath + "result/" + type + "result2.jpg").deleteOnExit();
242     }
243     /**
244      * 获取原始图url
245      * @param pageSource
246      * @return
247      */
248     private static String getFullImageUrl(String pageSource){
249         String url = null;
250         Document document = Jsoup.parse(pageSource);
251         String style = document.select("[class=gt_cut_fullbg_slice]").first().attr("style");
252         Pattern pattern = Pattern.compile("url\\(\"(.*)\"\\)");
253         Matcher matcher = pattern.matcher(style);
254         if (matcher.find()){
255             url = matcher.group(1);
256         }
257         url = url.replace(".webp", ".jpg");
258         System.out.println(url);
259         return url;
260     }
261     /**
262      * 获取带背景的url
263      * @param pageSource
264      * @return
265      */
266     private static String getBgImageUrl(String pageSource){
267         String url = null;
268         Document document = Jsoup.parse(pageSource);
269         String style = document.select(".gt_cut_bg_slice").first().attr("style");
270         Pattern pattern = Pattern.compile("url\\(\"(.*)\"\\)");
271         Matcher matcher = pattern.matcher(style);
272         if (matcher.find()){
273             url = matcher.group(1);
274         }
275         url = url.replace(".webp", ".jpg");
276         System.out.println(url);
277         return url;
278     }
279     public static boolean cutPic(String srcFile, String outFile, int x, int y,
280                                  int width, int height) {
281         FileInputStream is = null;
282         ImageInputStream iis = null;
283         try {
284             if (!new File(srcFile).exists()) {
285                 return false;
286             }
287             is = new FileInputStream(srcFile);
288             String ext = srcFile.substring(srcFile.lastIndexOf(".") + 1);
289             Iterator<ImageReader> it = ImageIO.getImageReadersByFormatName(ext);
290             ImageReader reader = it.next();
291             iis = ImageIO.createImageInputStream(is);
292             reader.setInput(iis, true);
293             ImageReadParam param = reader.getDefaultReadParam();
294             Rectangle rect = new Rectangle(x, y, width, height);
295             param.setSourceRegion(rect);
296             BufferedImage bi = reader.read(0, param);
297             File tempOutFile = new File(outFile);
298             if (!tempOutFile.exists()) {
299                 tempOutFile.mkdirs();
300             }
301             ImageIO.write(bi, ext, new File(outFile));
302             return true;
303         } catch (Exception e) {
304             e.printStackTrace();
305             return false;
306         } finally {
307             try {
308                 if (is != null) {
309                     is.close();
310                 }
311                 if (iis != null) {
312                     iis.close();
313                 }
314             } catch (IOException e) {
315                 e.printStackTrace();
316                 return false;
317             }
318         }
319     }
320     /**
321      * 图片拼接 （注意：必须两张图片长宽一致哦）
322      * @param files 要拼接的文件列表
323      * @param type  1横向拼接，2 纵向拼接
324      * @param targetFile 输出文件
325      */
326     private static void mergeImage(String[] files, int type, String targetFile) {
327         int length = files.length;
328         File[] src = new File[length];
329         BufferedImage[] images = new BufferedImage[length];
330         int[][] ImageArrays = new int[length][];
331         for (int i = 0; i < length; i++) {
332             try {
333                 src[i] = new File(files[i]);
334                 images[i] = ImageIO.read(src[i]);
335             } catch (Exception e) {
336                 throw new RuntimeException(e);
337             }
338             int width = images[i].getWidth();
339             int height = images[i].getHeight();
340             ImageArrays[i] = new int[width * height];
341             ImageArrays[i] = images[i].getRGB(0, 0, width, height, ImageArrays[i], 0, width);
342         }
343         int newHeight = 0;
344         int newWidth = 0;
345         for (int i = 0; i < images.length; i++) {
346             // 横向
347             if (type == 1) {
348                 newHeight = newHeight > images[i].getHeight() ? newHeight : images[i].getHeight();
349                 newWidth += images[i].getWidth();
350             } else if (type == 2) {// 纵向
351                 newWidth = newWidth > images[i].getWidth() ? newWidth : images[i].getWidth();
352                 newHeight += images[i].getHeight();
353             }
354         }
355         if (type == 1 && newWidth < 1) {
356             return;
357         }
358         if (type == 2 && newHeight < 1) {
359             return;
360         }
361         // 生成新图片
362         try {
363             BufferedImage ImageNew = new BufferedImage(newWidth, newHeight, BufferedImage.TYPE_INT_RGB);
364             int height_i = 0;
365             int width_i = 0;
366             for (int i = 0; i < images.length; i++) {
367                 if (type == 1) {
368                     ImageNew.setRGB(width_i, 0, images[i].getWidth(), newHeight, ImageArrays[i], 0,
369                             images[i].getWidth());
370                     width_i += images[i].getWidth();
371                 } else if (type == 2) {
372                     ImageNew.setRGB(0, height_i, newWidth, images[i].getHeight(), ImageArrays[i], 0, newWidth);
373                     height_i += images[i].getHeight();
374                 }
375             }
376             //输出想要的图片
377             ImageIO.write(ImageNew, targetFile.split("\\.")[1], new File(targetFile));
378 
379         } catch (Exception e) {
380             throw new RuntimeException(e);
381         }
382     }
383 }
384 
385 作者：卧颜沉默
386 链接：https://www.jianshu.com/p/1466f1ba3275
387 來源：简书
388 著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

pom文件依赖如下

 1     <dependency>
 2       <groupId>org.seleniumhq.selenium</groupId>
 3       <artifactId>selenium-server</artifactId>
 4       <version>3.0.1</version>
 5     </dependency>
 6     <!-- https://mvnrepository.com/artifact/org.jsoup/jsoup -->
 7     <dependency>
 8       <groupId>org.jsoup</groupId>
 9       <artifactId>jsoup</artifactId>
10       <version>1.7.2</version>
11     </dependency>
12 
13 作者：卧颜沉默
14 链接：https://www.jianshu.com/p/1466f1ba3275
15 來源：简书
16 著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。

最后

完整代码已上传至github,地址:https://github.com/wycm/selenium-geetest-crack
附上一张滑动效果图

刷新页面返回顶部

钱先生

selenium+java破解极验滑动验证码的示例代码

selenium+java破解极验滑动验证码

摘要

人工验证的过程

selenium模拟验证的过程

详细分析

最后

About