最近在做船只识别方面的事情，需要大量的正样本来训练adaboost分类器。于是到marinetraffic这个网站上下载船只图片。写个爬虫来自动下载显然很方便。
网站特点
在介绍爬虫之前首先了解一下marinetraffic这个网站的一些特点：
1. 会定期检测爬虫行为，如果认为有爬虫大量下载图片。会把该连接加入黑名单，后几天都没办法下载。
2. 船只图片资源差异大，有的船只有1000多张图，有的船只没有一张图，我们需要的是很多船只的很多张图，所以需要对下载的船只按优先级排序。
3. 用来训练分类器的正样本要求检测对象的分辨率一样，而marinetraffic网站下载的图片可以设置下在的图片的宽度，网站根据长宽比，生成相应的高度。所以，不同图片高度不一样，需要自己后期处理。
解决方案
- 针对爬虫检测，设置一个随机等待时间，10s左右。可以绕过网站爬虫行为检测。
- 对船只按照图片熟练排序，先下载图片数量多的，并且每个船只不用下载太多，保证图片的差异性。例如
- 在下载的时候使用统一的宽度。后期处理从图片中抠出分辨率一样的船只
爬虫源码
using System;using System.Collections.Generic;using System.Globalization;using System.IO;using System.Linq;using System.Net;using System.Runtime.Serialization.Formatters.Binary;using System.Text;using System.Text.RegularExpressions;using System.Threading;using System.Threading.Tasks;namespace 船只图像爬虫{ class Program { static void download_all_shipid(List<string> shipid_list) { try { WebClient MyWebClient = new WebClient(); MyWebClient.Headers["User-Agent"] = "blah"; MyWebClient.Credentials = CredentialCache.DefaultCredentials;//获取或设置用于向Internet资源的请求进行身份验证的网络凭据; //Console.WriteLine("here1"); //http://www.marinetraffic.com/en/photos/of/ships/shipid:281519/ //http://www.marinetraffic.com/en/ais/index/ships/all //http://www.marinetraffic.com/ais/index/ships/all/page:2/sort:COUNT_PHOTOS/direction:desc; for (int pageNum = 1; pageNum < 100; pageNum++) { Console.WriteLine("开始分析第" + pageNum + "张网页"); MyWebClient.Credentials = CredentialCache.DefaultCredentials;//获取或设置用于向Internet资源的请求进行身份验证的网络凭据; MyWebClient.Headers["User-Agent"] = "blah"; try { //Console.WriteLine("here0"); Byte[] pageData = MyWebClient.DownloadData(@"http://www.marinetraffic.com/en/ais/index/ships/all/page:" + pageNum + "/sort:COUNT_PHOTOS/direction:desc/per_page:50"); //从指定网站下载数据 //pageHtml = Encoding.Default.GetString(pageData); //如果获取网站页面采用的是GB2312，则使用这句; string pageHtml = Encoding.UTF8.GetString(pageData); //如果获取网站页面采用的是UTF-8，则使用这句; //Console.WriteLine(pageHtml);//在控制台输入获取的内容; //Console.WriteLine("here1"); int urlindex = -1; string org_label = "shipid:"; urlindex = pageHtml.IndexOf(org_label, urlindex + 1); while (urlindex != -1) { int endOfUrl = pageHtml.IndexOf("/", urlindex + org_label.Length); //Console.WriteLine("here2"); string shipid = pageHtml.Substring(urlindex + org_label.Length, endOfUrl - urlindex - org_label.Length); if (!shipid_list.Contains(shipid)) { Console.WriteLine("新增id:" + shipid); shipid_list.Add(shipid); } //Console.WriteLine("已有id:" + shipid); urlindex = pageHtml.IndexOf(org_label, urlindex + 1); } ///保存网页 //using (StreamWriter sw = new StreamWriter("ouput.html"))//将获取的内容写入文本 //{ // sw.Write(pageHtml); //} Console.WriteLine("完成第" + pageNum + "页分析"); } catch (WebException webEx) { Console.WriteLine(webEx.Message.ToString()); } //下面是一个随机数的方法保证10秒后再下载，以绕过违规检测。 Console.Write("绕开网站爬虫行为检测中......"); Random rd = new Random(); int time_sleep = rd.Next() % 10 + 10; Thread.Sleep(time_sleep * 1000); Console.WriteLine(); } Console.WriteLine("分析结束"); //下面把list内容保存进文件,使用序列化的方法; string file = @"C:\Users\dragonfive\Desktop\爬虫获得船只图片\第三批\0_100page_shipid.txt"; using (FileStream fsWriter = new FileStream(file, FileMode.OpenOrCreate, FileAccess.Write)) { //下面对stu进行序列化； BinaryFormatter bf = new BinaryFormatter(); bf.Serialize(fsWriter, shipid_list); } } catch (WebException webEx) { Console.WriteLine(webEx.Message.ToString()); } } /// <summary> /// 根据得到的ship_id获得该ship_id的所有图片; /// </summary> /// <param name="ship_id"></param> static void download_jpg(string ship_id) { try { Console.WriteLine("开始下载shipid为："+ship_id+"的图片"); WebClient MyWebClient = new WebClient(); MyWebClient.Credentials = CredentialCache.DefaultCredentials;//获取或设置用于向Internet资源的请求进行身份验证的网络凭据 MyWebClient.Headers["User-Agent"] = "blah"; //http://www.marinetraffic.com/en/photos/of/ships/shipid:281519/ //http://www.marinetraffic.com/en/photos/of/ships/shipid:371668/per_page:1000/page:1 Byte[] pageData = MyWebClient.DownloadData(@"http://www.marinetraffic.com/en/photos/of/ships/shipid:" + ship_id + @"/per_page:100/page:1"); //从指定网站下载数据 //string pageHtml = Encoding.Default.GetString(pageData); //如果获取网站页面采用的是GB2312，则使用这句 string pageHtml = Encoding.UTF8.GetString(pageData); //如果获取网站页面采用的是UTF-8，则使用这句 //Console.WriteLine(pageHtml);//在控制台输入获取的内容 Console.WriteLine("元网页已下载"); //using (StreamWriter sw = new StreamWriter("ouput.html"))//将获取的内容写入文本 //{ // sw.Write(pageHtml); //} int urlindex = -1; string org_label = "data-original='"; urlindex = pageHtml.IndexOf(org_label, urlindex + 1); int i = 0; //Directory.CreateDirectory(@"./" ); while (urlindex != -1) { int endOfUrl = pageHtml.IndexOf("'", urlindex + org_label.Length); string url = pageHtml.Substring(urlindex + org_label.Length, endOfUrl - urlindex - org_label.Length); ////下面是unicode编码转换为string的方式; //MatchCollection mc = Regex.Matches(strName, @"\\u([\w]{2})([\w]{2})", RegexOptions.Compiled | RegexOptions.IgnoreCase); //byte[] bts = new byte[2]; //foreach (Match m in mc) //{ // bts[0] = (byte)int.Parse(m.Groups[2].Value, NumberStyles.HexNumber); // bts[1] = (byte)int.Parse(m.Groups[1].Value, NumberStyles.HexNumber); // musicName += Encoding.Unicode.GetString(bts); //} //Console.WriteLine("接下来下载的是:" + musicName); //下面是一个随机数的方法保证10秒后再下载，以绕过违规检测。 Console.Write("绕过网站爬虫行为检测中......"); Random rd = new Random(); int time_sleep = rd.Next() % 10 + 10; Thread.Sleep(time_sleep * 1000); Console.WriteLine(); try { //这是下载的命令； Console.WriteLine(url); MyWebClient.Credentials = CredentialCache.DefaultCredentials;//获取或设置用于向Internet资源的请求进行身份验证的网络凭据 MyWebClient.Headers["User-Agent"] = "blah"; Byte[] jpgdata = MyWebClient.DownloadData(url); //从指定网页下载数据; //把下载的内容保存在一个地方; using (FileStream fs = new FileStream(@"C:\Users\dragonfive\Desktop\爬虫获得船只图片\第三批\" + ship_id + "_" + i + ".jpg", FileMode.OpenOrCreate, FileAccess.Write)) { fs.Write(jpgdata, 0, jpgdata.Length); } } catch (WebException webEx) { Console.WriteLine("被捕获了吗?"); Console.WriteLine(webEx.Message.ToString()); } Console.WriteLine("成功下载第" + (i ++) + "张图片"); urlindex = pageHtml.IndexOf(org_label, urlindex + 1); } ///保存网页 //using (StreamWriter sw = new StreamWriter("ouput.html"))//将获取的内容写入文本 //{ // sw.Write(pageHtml); //} Console.WriteLine("*****************************************"); Console.WriteLine("下载"+i+"张ship_id为"+ship_id+"的图片"); Console.WriteLine("*****************************************"); //Console.ReadLine(); //让控制台暂停,否则一闪而过了 } catch (WebException webEx) { Console.WriteLine(webEx.Message.ToString()); } } static void Main(string[] args) { List<string> shipid_list = new List<string>(); //shipid_list.Add("371681");//暂时快速产生图片用这个; download_all_shipid(shipid_list); //string file = @"C:\Users\dragonfive\Desktop\爬虫获得船只图片\第三批\0_100page_shipid.txt"; //using (FileStream fsReader = new FileStream(file, FileMode.Open, FileAccess.Read)) //{ // //下面进行反序列话; // BinaryFormatter bf = new BinaryFormatter(); // shipid_list = (List<string>)bf.Deserialize(fsReader); // Console.WriteLine("成功载入" + shipid_list.Count + "个shipid"); //} ////371652 371668 371681 1252401 //shipid_list.Remove("371652"); //shipid_list.Remove("371668"); //shipid_list.Remove("371681"); //shipid_list.Remove("1252401"); ////132264 //shipid_list.Remove("371077"); //shipid_list.Remove("132264"); //shipid_list.Remove("224871"); //shipid_list.Remove("279923"); //shipid_list.Remove("369163"); //shipid_list.Remove("266342"); //shipid_list.Remove("371216"); //shipid_list.Remove("368174"); //shipid_list.Remove("369163"); foreach (var ship_id in shipid_list) { download_jpg(ship_id); } Console.ReadLine(); //让控制台暂停,否则一闪而过了 } }}版权声明：本文为博主原创文章，欢迎转载和分享，但请声明出处http://blog.csdn.net/zhzz2012