利用java抓取网页某个图片

import java.io.File;
import java.io.FileOutputStream;
import java.io.InputStream;
import java.net.URL;
import java.net.URLConnection;
import java.util.ArrayList;
import java.util.List;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

public class CatchImage {
 public static void main(String[] args) throws Exception { 
        CatchImage cm = new CatchImage(); 
         //获得html文本内容  
         String HTML = cm.getHTML(URL); 
         //获取图片标签  
          List<String> imgUrl = cm.getImageUrl(HTML); 
          //获取图片src地址  
         List<String> imgSrc = cm.getImageSrc(imgUrl); 
          //下载图片  
          cm.Download(imgSrc); 
         System.out.println("抓取的图片信息存放在:"+System.getProperty("user.dir"));

      } 
   /**
    *
    * @param url
    * @return
    * @throws Exception
    */
  private String getHTML(String url) throws Exception {
         URL uri = new URL(url); 
         URLConnection connection = uri.openConnection(); 
         InputStream in = connection.getInputStream(); 
         byte[] buf = new byte[1024]; 
         int length = 0;
         StringBuffer sb = new StringBuffer(); 
        while ((length = in.read(buf, 0, buf.length)) > 0) { 
             sb.append(new String(buf, ECODING)); 
         } 
         in.close(); 
         return sb.toString(); 
    } 
    
     /***
          * 获取ImageUrl地址
          * 
          * @param HTML
          * @return
         */ 
         private List<String> getImageUrl(String HTML) { 
            Matcher matcher = Pattern.compile(IMGURL_REG).matcher(HTML); 
             List<String> listImgUrl = new ArrayList<String>(); 
             while (matcher.find()) { 
                 listImgUrl.add(matcher.group()); 
             } 
             return listImgUrl; 
         } 
      
         /***
          * 获取ImageSrc地址
          * 
          * @param listImageUrl
          * @return
          */ 
         private List<String> getImageSrc(List<String> listImageUrl) { 
             List<String> listImgSrc = new ArrayList<String>(); 
             for (String image : listImageUrl) { 
                 Matcher matcher = Pattern.compile(IMGSRC_REG).matcher(image); 
                 while (matcher.find()) { 
                     listImgSrc.add("http://www.songjiangaqi.com.cn/home/"+matcher.group().substring(5, matcher.group().length())); 
                 } 
             } 
            return listImgSrc; 
         } 
         /***
             * 下载图片
             * 
             * @param listImgSrc
              */ 
             private void Download(List<String> listImgSrc) { 
                 try { 
                     for (String url : listImgSrc) { 
                         String imageName = url.substring(url.lastIndexOf("/") + 1, url.length()); 
                         URL uri = new URL(url); 
                         InputStream in = uri.openStream(); 
                         FileOutputStream fo = new FileOutputStream(new File(imageName)); 
                         byte[] buf = new byte[1024]; 
                         int length = 0; 
                         System.out.println("开始下载:" + url); 
                         while ((length = in.read(buf, 0, buf.length)) != -1) { 
                             fo.write(buf, 0, length); 
                        } 
                         in.close(); 
                         fo.close(); 
                         System.out.println(imageName + "下载完成"); 
                     } 
                 } catch (Exception e) { 
                     System.out.println("下载失败"); 
                 } 
             } 

         // 地址  
            private static final String URL = "http://www.songjiangaqi.com.cn/home/index.aspx";
           // 编码  
            private static final String ECODING = "UTF-8"; 
            // 获取img标签正则  
               private static final String IMGURL_REG = "<img id=\"Image1\".*src=(.*?)[^>]*?>"; 
             // 获取src路径的正则  
             private static final String IMGSRC_REG = "src=(.*?)png"; 

}