转自:http://blog.csdn.net/augusdi/article/details/9012043
otsu法(最大类间方差法,有时也称之为大津算法)使用的是聚类的思想,把图像的灰度数按灰度级分成2个部分,使得两个部分之间的灰度值差异最大,每个部分之间的灰度差异最小,通过方差的计算来寻找一个合适的灰度级别 来划分。 所以 可以在二值化的时候 采用otsu算法来自动选取阈值进行二值化。otsu算法被认为是图像分割中阈值选取的最佳算法,计算简单,不受图像亮度和对比度的影响。因此,使类间方差最大的分割意味着错分概率最小。
设t为设定的阈值。
wo: 分开后 前景像素点数占图像的比例
uo: 分开后 前景像素点的平均灰度
w1:分开后 被景像素点数占图像的比例
u1: 分开后 被景像素点的平均灰度
u=w0*u0 + w1*u1 :图像总平均灰度
从L个灰度级遍历t,使得t为某个值的时候,前景和背景的方差最大, 则 这个 t 值便是我们要求得的阈值。
其中,方差的计算公式如下:
g=wo * (uo - u) * (uo - u) + w1 * (u1 - u) * (u1 - u)
[ 此公式计算量较大,可以采用: g = wo * w1 * (uo - u1) * (uo - u1) ]
由于otsu算法是对图像的灰度级进行聚类,so 在执行otsu算法之前,需要计算该图像的灰度直方图。
按照上面的解释参考代码如下:
- #include <stdio.h>
- #include <cv.h>
- #include <highgui.h>
- #include <math.h>
- #pragma comment(lib, "cv.lib")
- #pragma comment(lib, "cxcore.lib")
- #pragma comment(lib, "highgui.lib")
- int Otsu(IplImage* src);
- int _tmain(int argc, _TCHAR* argv[])
- {
- IplImage* img = cvLoadImage("c:\\aSa.jpg",0);
- IplImage* dst = cvCreateImage(cvGetSize(img), 8, 1);
- int threshold = Otsu(img);
- cvThreshold(img, dst, threshold, 255, CV_THRESH_BINARY);
- cvNamedWindow( "img", 1 );
- cvShowImage("img", dst);
- cvWaitKey(-1);
- cvReleaseImage(&img);
- cvReleaseImage(&dst);
- cvDestroyWindow( "dst" );
- return 0;
- }
- int Otsu(IplImage* src)
- {
- int height=src->height;
- int width=src->width;
- long size = height * width;
- //histogram
- float histogram[256] = {0};
- for(int m=0; m < height; m++)
- {
- unsigned char* p=(unsignedchar*)src->imageData + src->widthStep * m;
- for(int n = 0; n < width; n++)
- {
- histogram[int(*p++)]++;
- }
- }
- int threshold;
- long sum0 = 0, sum1 = 0;//存储前景的灰度总和和背景灰度总和
- long cnt0 = 0, cnt1 = 0; //前景的总个数和背景的总个数
- double w0 = 0, w1 = 0;//前景和背景所占整幅图像的比例
- double u0 = 0, u1 = 0; //前景和背景的平均灰度
- double variance = 0;//最大类间方差
- int i, j;
- double u = 0;
- double maxVariance = 0;
- for(i = 1; i < 256; i++)//一次遍历每个像素
- {
- sum0 = 0;
- sum1 = 0;
- cnt0 = 0;
- cnt1 = 0;
- w0 = 0;
- w1 = 0;
- for(j = 0; j < i; j++)
- {
- cnt0 += histogram[j];
- sum0 += j * histogram[j];
- }
- u0 = (double)sum0 / cnt0;
- w0 = (double)cnt0 / size;
- for(j = i ; j <= 255; j++)
- {
- cnt1 += histogram[j];
- sum1 += j * histogram[j];
- }
- u1 = (double)sum1 / cnt1;
- w1 = 1 - w0; // (double)cnt1 / size;
- u = u0 * w0 + u1 * w1; //图像的平均灰度
- printf("u = %f\n", u);
- //variance = w0 * pow((u0 - u), 2) + w1 * pow((u1 - u), 2);
- variance = w0 * w1 * (u0 - u1) * (u0 - u1);
- if(variance > maxVariance)
- {
- maxVariance = variance;
- threshold = i;
- }
- }
- printf("threshold = %d\n", threshold);
- return threshold;
- }
#include <stdio.h>
#include <cv.h>
#include <highgui.h>
#include <math.h>
#pragma comment(lib, "cv.lib")
#pragma comment(lib, "cxcore.lib")
#pragma comment(lib, "highgui.lib")
int Otsu(IplImage* src);
int _tmain(int argc, _TCHAR* argv[])
{
IplImage* img = cvLoadImage("c:\\aSa.jpg",0);
IplImage* dst = cvCreateImage(cvGetSize(img), 8, 1);
int threshold = Otsu(img);
cvThreshold(img, dst, threshold, 255, CV_THRESH_BINARY);
cvNamedWindow( "img", 1 );
cvShowImage("img", dst);
cvWaitKey(-1);
cvReleaseImage(&img);
cvReleaseImage(&dst);
cvDestroyWindow( "dst" );
return 0;
}
int Otsu(IplImage* src)
{
int height=src->height;
int width=src->width;
long size = height * width;
//histogram
float histogram[256] = {0};
for(int m=0; m < height; m++)
{
unsigned char* p=(unsigned char*)src->imageData + src->widthStep * m;
for(int n = 0; n < width; n++)
{
histogram[int(*p++)]++;
}
}
int threshold;
long sum0 = 0, sum1 = 0; //存储前景的灰度总和和背景灰度总和
long cnt0 = 0, cnt1 = 0; //前景的总个数和背景的总个数
double w0 = 0, w1 = 0; //前景和背景所占整幅图像的比例
double u0 = 0, u1 = 0; //前景和背景的平均灰度
double variance = 0; //最大类间方差
int i, j;
double u = 0;
double maxVariance = 0;
for(i = 1; i < 256; i++) //一次遍历每个像素
{
sum0 = 0;
sum1 = 0;
cnt0 = 0;
cnt1 = 0;
w0 = 0;
w1 = 0;
for(j = 0; j < i; j++)
{
cnt0 += histogram[j];
sum0 += j * histogram[j];
}
u0 = (double)sum0 / cnt0;
w0 = (double)cnt0 / size;
for(j = i ; j <= 255; j++)
{
cnt1 += histogram[j];
sum1 += j * histogram[j];
}
u1 = (double)sum1 / cnt1;
w1 = 1 - w0; // (double)cnt1 / size;
u = u0 * w0 + u1 * w1; //图像的平均灰度
printf("u = %f\n", u);
//variance = w0 * pow((u0 - u), 2) + w1 * pow((u1 - u), 2);
variance = w0 * w1 * (u0 - u1) * (u0 - u1);
if(variance > maxVariance)
{
maxVariance = variance;
threshold = i;
}
}
printf("threshold = %d\n", threshold);
return threshold;
}
把w1写成w0 ··害我debug 了好久~~总是不认真,脑袋浑浑噩噩的···这都看不出来。。。。
==================
对了,之前搜集的一个otsu的算法,代码如下(由于时间太久了,不知道出处了。。。膜拜大牛哈)
- <p>#include <stdio.h>
- #include <cv.h>
- #include <highgui.h>
- #include <math.h></p><p>#pragma comment(lib, "cv.lib")
- #pragma comment(lib, "cxcore.lib")
- #pragma comment(lib, "highgui.lib")
- </p>int Otsu(IplImage* src);
- int _tmain(int argc, _TCHAR* argv[])
- {
- IplImage* img = cvLoadImage("c:\\aSa.jpg",0);
- IplImage* dst = cvCreateImage(cvGetSize(img), 8, 1);
- int threshold = Otsu(img);
- printf("threshold = %d\n", threshold);
- cvThreshold(img, dst, threshold, 255, CV_THRESH_BINARY);
- cvNamedWindow( "img", 1 );
- cvShowImage("img", dst);
- cvWaitKey(-1);
- cvReleaseImage(&img);
- cvReleaseImage(&dst);
- cvDestroyWindow( "dst" );
- return 0;
- }
- int Otsu(IplImage* src)
- {
- int height=src->height;
- int width=src->width;
- //histogram
- float histogram[256] = {0};
- for(int i=0; i < height; i++)
- {
- unsigned char* p=(unsignedchar*)src->imageData + src->widthStep * i;
- for(int j = 0; j < width; j++)
- {
- histogram[*p++]++;
- }
- }
- //normalize histogram
- int size = height * width;
- for(int i = 0; i < 256; i++)
- {
- histogram[i] = histogram[i] / size;
- }
- //average pixel value
- float avgValue=0;
- for(int i=0; i < 256; i++)
- {
- avgValue += i * histogram[i]; //整幅图像的平均灰度
- }
- int threshold;
- float maxVariance=0;
- float w = 0, u = 0;
- for(int i = 0; i < 256; i++)
- {
- w += histogram[i]; //假设当前灰度i为阈值, 0~i 灰度的像素(假设像素值在此范围的像素叫做前景像素) 所占整幅图像的比例
- u += i * histogram[i]; // 灰度i 之前的像素(0~i)的平均灰度值: 前景像素的平均灰度值
- float t = avgValue * w - u;
- float variance = t * t / (w * (1 - w) );
- if(variance > maxVariance)
- {
- maxVariance = variance;
- threshold = i;
- }
- }
- return threshold;
- }