OTSU 是一个自动的阈值分割
主要针对全局分割,是个自适应二分类问题
记T为前景与背景的分割阈值,前景点数占图像比例为ω0,平均灰度为μ0;背景点数占图像比例为ω1,平均灰度为μ1,图像的总平均灰度为μ,前景和背景图象的方差g,则
ω0=N0/ M×N (1)
ω1=N1/ M×N (2)
N0+N1=M×N (3)
ω0+ω1=1 (4)
μ=ω0*μ0+ω1*μ1 (5)
g=ω0(μ0-μ)^2+ω1(μ1-μ)^2 (6)
将式(4)(5)代入式(6),得到等价公式:
g=ω0ω1(μ0-μ1)^2 (7) 这就是类间方差
所以,只要对于阈值T从0-255遍历一次,就能求得对应的g,将g最大时的T值记录下来即可
opencv中的实现如下
static double
getThreshVal_Otsu_8u( const Mat& _src )
{
Size size = _src.size();
int step = (int) _src.step;
if( _src.isContinuous() )
{
size.width *= size.height;
size.height = 1;
step = size.width;
}
#ifdef HAVE_IPP
unsigned char thresh = 0;
CV_IPP_RUN_FAST(ipp_getThreshVal_Otsu_8u(_src.ptr(), step, size, thresh), thresh);
#endif
const int N = 256;
int i, j, h[N] = {0};
for( i = 0; i < size.height; i++ )
{
const uchar* src = _src.ptr() + step*i;
j = 0;
#if CV_ENABLE_UNROLLED
for( ; j <= size.width - 4; j += 4 )
{
int v0 = src[j], v1 = src[j+1];
h[v0]++; h[v1]++;
v0 = src[j+2]; v1 = src[j+3];
h[v0]++; h[v1]++;
}
#endif
for( ; j < size.width; j++ )
h[src[j]]++;
}
double mu = 0, scale = 1./(size.width*size.height);
for( i = 0; i < N; i++ )
mu += i*(double)h[i];
mu *= scale;
double mu1 = 0, q1 = 0;
double max_sigma = 0, max_val = 0;
for( i = 0; i < N; i++ )
{
double p_i, q2, mu2, sigma;
p_i = h[i]*scale;
mu1 *= q1;
q1 += p_i;
q2 = 1. - q1;
if( std::min(q1,q2) < FLT_EPSILON || std::max(q1,q2) > 1. - FLT_EPSILON )
continue;
mu1 = (mu1 + i*p_i)/q1;
mu2 = (mu - q1*mu1)/q2;
sigma = q1*q2*(mu1 - mu2)*(mu1 - mu2);
if( sigma > max_sigma )
{
max_sigma = sigma;
max_val = i;
}
}
return max_val;
}
opencv中的OTSU除法有点多,其实可以在循环中优化(使用前一次的累加)