使用C#的aforge类库识别验证码实例 - 黑马程序员技术交流社区

//转化图片像素格式
var bnew = new Bitmap(b.Width, b.Height, PixelFormat.Format24bppRgb);
Graphics g = Graphics.FromImage(bnew);
g.DrawImage(b, 0, 0);
g.Dispose();

复制代码

//灰度化
b = new Grayscale(0.2125, 0.7154, 0.0721).Apply(b);

复制代码

//二值化
b = new Threshold(50).Apply(b);

复制代码

//去噪点
new BlobsFiltering(1, 1, b.Width, b.Height).Apply(b);

复制代码

/// <summary>
/// 按照 Y 轴线切割
/// (丢弃等于号)
/// </summary>
/// <param name="?"></param>
/// <returns></returns>
public List<Bitmap> Crop_Y(Bitmap b)
{
var list = new List<Bitmap>();
//统计每一列的“1”的个数，方便切除
int[] cols = new int[b.Width];
/*
* 纵向切割
*/
for (int x = 0; x < b.Width; x++)
{
for (int y = 0; y < b.Height; y++)
{
//获取当前像素点像素
var pixel = b.GetPixel(x, y);
//说明是黑色点
if (pixel.R == 0)
{
cols[x] = ++cols[x];
}
}
}
int left = 0, right = 0;
for (int i = 0; i < cols.Length; i++)
{
//说明该列有像素值（为了防止像素干扰，去噪后出现空白的问题，所以多判断一下，防止切割成多个)
if (cols[i] > 0 || (i + 1 < cols.Length && cols[i + 1] > 0))
{
if (left == 0)
{
//切下来图片的横坐标left
left = i;
}
else
{
//切下来图片的横坐标right
right = i;
}
}
else
{
//说明已经有切割图了，下面我们进行切割处理
if ((left > 0 || right > 0))
{
Crop corp = new Crop(new Rectangle(left, 0, right - left + 1, b.Height));
var small = corp.Apply(b);
//居中，将图片放在20*50的像素里面
list.Add(small);
}
left = right = 0;
}
}
return list;
}
/// <summary>
/// 按照 X 轴线切割
/// </summary>
/// <param name="b"></param>
/// <returns></returns>
public List<Bitmap> Crop_X(List<Bitmap> list)
{
var corplist = new List<Bitmap>();
//再对分割的图进行上下切割，取出上下的白边
foreach (var segb in list)
{
//统计每一行的“1”的个数，方便切除
int[] rows = new int[segb.Height];
/*
* 横向切割
*/
for (int y = 0; y < segb.Height; y++)
{
for (int x = 0; x < segb.Width; x++)
{
//获取当前像素点像素
var pixel = segb.GetPixel(x, y);
//说明是黑色点
if (pixel.R == 0)
{
rows[y] = ++rows[y];
}
}
}
int bottom = 0, top = 0;
for (int y = 0; y < rows.Length; y++)
{
//说明该行有像素值（为了防止像素干扰，去噪后出现空白的问题，所以多判断一下，防止切割成多个)
if (rows[y] > 0 || (y + 1 < rows.Length && rows[y + 1] > 0))
{
if (top == 0)
{
//切下来图片的top坐标
top = y;
}
else
{
//切下来图片的bottom坐标
bottom = y;
}
}
else
{
//说明已经有切割图了，下面我们进行切割处理
if ((top > 0 || bottom > 0) && bottom - top > 0)
{
Crop corp = new Crop(new Rectangle(0, top, segb.Width, bottom - top + 1));
var small = corp.Apply(segb);
corplist.Add(small);
}
top = bottom = 0;
}
}
}
return corplist;
}

复制代码

/// <summary>
/// 重置图片的指定大小并且居中
/// </summary>
/// <param name="list"></param>
/// <returns></returns>
public List<Bitmap> ToResizeAndCenterIt(List<Bitmap> list, int w = 20, int h = 20)
{
List<Bitmap> resizeList = new List<Bitmap>();
for (int i = 0; i < list.Count; i++)
{
//反转一下图片
list[i] = new Invert().Apply(list[i]);
int sw = list[i].Width;
int sh = list[i].Height;
Crop corpFilter = new Crop(new Rectangle(0, 0, w, h));
list[i] = corpFilter.Apply(list[i]);
//再反转回去
list[i] = new Invert().Apply(list[i]);
//计算中心位置
int centerX = (w - sw) / 2;
int centerY = (h - sh) / 2;
list[i] = new CanvasMove(new IntPoint(centerX, centerY), Color.White).Apply(list[i]);
resizeList.Add(list[i]);
}
return resizeList;
}

复制代码

ExhaustiveTemplateMatching templateMatching = new ExhaustiveTemplateMatching(0.9f);

复制代码

var files = Directory.GetFiles(Environment.CurrentDirectory + "\\Template\\");
var templateList = files.Select(i => { return new Bitmap(i); }).ToList();
var templateListFileName = files.Select(i => { return i.Substring(30, 1); }).ToList();
var result = new List<string>();
ExhaustiveTemplateMatching templateMatching = new ExhaustiveTemplateMatching(0.9f);
//这里面有四张图片，进行四张图的模板匹配
for (int i = 0; i < list.Count; i++)
{
float max = 0;
int index = 0;
for (int j = 0; j < templateList.Count; j++)
{
var compare = templateMatching.ProcessImage(list[i], templateList[j]);
if (compare.Length > 0 && compare[0].Similarity > max)
{
//记录下最相似的
max = compare[0].Similarity;
index = j;
}
}
result.Add(templateListFileName[index]);
}

复制代码