区域对比验证码识别，超级简单的Region对比识别验证码

 在验证识别常用技巧之外，我们可以采用区域对比的方法进行验证码识别，对相对简单的验证码有效，如果由噪点的话可以现去噪点，然后再执行如下函数

 

在验证识别常用技巧之外，我们可以采用区域对比的方法进行验证码识别，对相对简单的验证码有效，如果由噪点的话可以现去噪点，然后再执行如下函数

在一些验证码相对简单的站点中，就是对于特定的字符，在生成时，其字符的Region应该是一样的，如图片



中的6总是这个字体这个字号......我们只须把验证码图片从左向右一列一列扫描，分隔出每个字符的RGN，然后和chars.bmp中的每个字符的RGN对比，就知道是哪个数字了。

 　　关键代码如下：

namespace WindowsApplication1
{
public partial class Form1 : Form
{
IniFile config = null;

//查看两个颜色是不是一样，注意这里有一定误差也算相同
public bool IsSameColor(Color c1, Color c2)
{
if (Math.Abs(c1.R - c2.R) < 10
&& Math.Abs(c1.G - c2.G) < 10
&& Math.Abs(c1.B - c2.B) < 10)
{
return true;
}
else
{
return false;
}
}

//计算一个Region中，像素的个数
public int RegionPointCount(Region r,int width,int height)
{
int count = 0;
for(int h = 0; h<width; ++h)
{
for(int v = 0; v< height; ++v)
{
if(r.IsVisible(h,v))
{
++count;
}
}
}

return count;
}

//初始化每个字符的Region
public void InitPictureCharInfo()
{
Bitmap bmp = (Bitmap)Bitmap.FromFile("chars.bmp");
List<BitmapCharInfo> bcil = charRgnList;
Region rgn = new Region();
rgn.MakeEmpty();
if (bmp.Height > 0 && bmp.Width > 0)
{
Color bkColor = bmp.GetPixel(0, 0);
bool bInWorking = false;
int nNextStartPos = 0;
for (int h = 0; h < bmp.Width; ++h)
{
bool bFindColor = false;
for (int v = 0; v < bmp.Height; ++v)
{
if (!IsSameColor(bkColor, bmp.GetPixel(h, v)))
{
rgn.Union(new Rectangle(h, v, 1, 1));
bFindColor = true;
}
}

if (bInWorking)
{
if (!bFindColor)
{
bInWorking = false;
rgn.Translate(-nNextStartPos, 0);
BitmapCharInfo bci = new BitmapCharInfo(rgn, h - nNextStartPos, bmp.Height);
bci.orgPos = nNextStartPos;
bcil.Add(bci);
rgn = new Region();
rgn.MakeEmpty();
}
}
else
{
if (bFindColor)
{
bInWorking = true;
nNextStartPos = h;
}
}
}

chars.AddRange("0123456789".ToCharArray());
}
}

//扫描并识别验证码
public void ScanValidCode()
{
Bitmap bmp = this.bmpValidCode;
List<BitmapCharInfo> bcil = new List<BitmapCharInfo>();
Region rgn = new Region();
rgn.MakeEmpty();
if (bmp.Height > 0 && bmp.Width > 0)
{
Color bkColor = bmp.GetPixel(0, 0);
bool bInWorking = false;
int nNextStartPos = 0;
for (int h = 0; h < bmp.Width; ++h)
{
bool bFindColor = false;
for (int v = 0; v < bmp.Height; ++v)
{
if (!IsSameColor(bkColor, bmp.GetPixel(h, v)))
{
rgn.Union(new Rectangle(h, v, 1, 1));
bFindColor = true;
}
}

if (bInWorking)
{
if (!bFindColor)
{
bInWorking = false;
rgn.Translate(-nNextStartPos, 0);
BitmapCharInfo bci = new BitmapCharInfo(rgn, h - nNextStartPos, bmp.Height);
bci.orgPos = nNextStartPos;
bcil.Add(bci);
rgn = new Region();
rgn.MakeEmpty();
}
}
else
{
if (bFindColor)
{
bInWorking = true;
nNextStartPos = h;
}
}
}

List<char> chs = new List<char>();

Graphics gh = Graphics.FromImage(bmp);
foreach (BitmapCharInfo bci in bcil)
{
int minPos = -1;
int minLng = -1;
for (int i = 0; i < charRgnList.Count; ++i)
{
Region r = bci.rgn.Clone();
r.Union(charRgnList[i].rgn);
r.Exclude(bci.rgn);

int lng = RegionPointCount(r, bci.width, bci.height);

if (minLng == -1)
{
minLng = lng;
minPos = i;
}
else
{
if (lng < minLng)
{
minLng = lng;
minPos = i;
}
}

}

if (minPos != -1)
{
chs.Add(chars[minPos]);
}
}

string str = new string(chs.ToArray(), 0, chs.Count);
//MessageBox.Show(str);
this.currScanValidCode = str;

}
}

public void GetValidCodePicture(CookieContainer cc, WebProxy wp )
{
//this.pbValidCode.ImageLocation = this.pbValidCode.ImageLocation;
Exception exp = null;
for (int i = 0; i < 3; ++i)
{
try
{
Stream s = GetDataNonProxy(this.tbValidPicUrl.Text, cc,wp);
Bitmap bmp = (Bitmap)Bitmap.FromStream(s);
s.Close();

bmpValidCode = bmp;

return;
}
catch (Exception ex)
{
exp = ex;
}
}

MessageBox.Show("猎取图片失败:" + exp == null?"unkown":exp.Message);
}

Bitmap bmpValidCode = null;
private void btnLoadPic_Click(object sender, EventArgs e)
{

this.GetValidCodePicture(null,null);
currValidCode = this.currScanValidCode;
this.UpdateDate(false);
}

List<BitmapCharInfo> charRgnList = new List<BitmapCharInfo>();
List<char> chars = new List<char>();
private void btnInit_Click(object sender, EventArgs e)
{
InitPictureCharInfo();
}

private void btnScanValidCode_Click(object sender, EventArgs e)
{
ScanValidCode();
this.currValidCode = this.currScanValidCode;
this.UpdateDate(false);

}

string currScanValidCode;

private void Form1_Load(object sender, EventArgs e)
{
this.UpdateDate(true);
InitPictureCharInfo();
config = new IniFile(Application.ExecutablePath + ".ini");
lastPwdPosInDict = config.GetInt("Process", "PwdPostion", 0);
InitPwdDict();
UpdateProcessText();
}

}

public class BitmapCharInfo
{
public BitmapCharInfo()
{
this.rgn = new Region();
this.rgn.MakeEmpty();
}
public BitmapCharInfo(Region r, int w, int h)
{
this.rgn = r;
this.width = w;
this.height = h;
}
public Region rgn;
public int width = 0;
public int height = 0;
public int orgPos = 0;

}
}

　识别效率提高的，但是局限性太高了。现在这类我网站也少得多了，现在放出来。希望抛砖引玉引出图形的高级算法来。