OpenCV混合高斯模型函数注释说明
一、cvaux.h
#define CV_BGFG_MOG_MAX_NGAUSSIANS 500
//高斯背景检测算法的默认参数设置
#define CV_BGFG_MOG_BACKGROUND_THRESHOLD 0.7 //高斯分布权重之和阈值
#define CV_BGFG_MOG_STD_THRESHOLD 2.5 //λ=2.5（99%）
#define CV_BGFG_MOG_WINDOW_SIZE 200 //学习率α=1/win_size
#define CV_BGFG_MOG_NGAUSSIANS 5 //k=5个混合高斯模型
#define CV_BGFG_MOG_WEIGHT_INIT 0.05 //初始权重
#define CV_BGFG_MOG_SIGMA_INIT 30 //初始标准差
#define CV_BGFG_MOG_MINAREA 15.f //？？？
#define CV_BGFG_MOG_NCOLORS 3 //颜色通道数
/************* CV_BG_STAT_MODEL_FIELDS()的宏定义**********************/
#define CV_BG_STAT_MODEL_FIELDS()
int type; //type of BG model
CvReleaseBGStatModel release; //
CvUpdateBGStatModel update;
IplImage* background; /*8UC3 reference background image*/
IplImage* foreground; /*8UC1 foreground image*/
IplImage** layers; /*8UC3 reference background image, can be null */
int layer_count; /* can be zero */
CvMemStorage* storage; /*storage for foreground_regions?/
CvSeq* foreground_regions /*foreground object contours*/
/*************************高斯背景模型参数结构体*************************/
typedef struct CvGaussBGStatModelParams
{
int win_size; //等于 1/alpha
int n_gauss; //高斯模型的个数
double bg_threshold, std_threshold, minArea; // bg_threshold：高斯分布权重之和阈值、std_threshold：2.5、minArea：？？？
double weight_init, variance_init; //权重和方差
}CvGaussBGStatModelParams;
/**************************高斯分布模型结构体***************************/
typedef struct CvGaussBGValues
{
int match_sum;
double weight;
double variance[CV_BGFG_MOG_NCOLORS];
double mean[CV_BGFG_MOG_NCOLORS];
}
CvGaussBGValues;
typedef struct CvGaussBGPoint
{
CvGaussBGValues* g_values;
}
CvGaussBGPoint;
/*************************高斯背景模型结构体*************************/
typedef struct CvGaussBGModel
{
CV_BG_STAT_MODEL_FIELDS();
CvGaussBGStatModelParams params;
CvGaussBGPoint* g_point;
int countFrames;
}
CvGaussBGModel;
二、cvbgfg_gaussmix.cpp
//////////////////////////////////////////////////////////// cvCreateGaussianBGModel////////////////////////////////////////////////////////////////
功能：高斯背景模型变量bg_model初始化赋值
CV_IMPL CvBGStatModel* cvCreateGaussianBGModel( IplImage* first_frame, CvGaussBGStatModelParams* parameters)
{
CvGaussBGModel* bg_model = 0; //高斯背景状态模型变量
CV_FUNCNAME( "cvCreateGaussianBGModel" );
__BEGIN__;
double var_init;
CvGaussBGStatModelParams params; //高斯背景状态模型参数变量
int i, j, k, n, m, p;
//初始化参数，如果参数为空，则进行初始化赋值
if( parameters == NULL )
{
params.win_size = CV_BGFG_MOG_WINDOW_SIZE; //学习率α=1/200=0.005
params.bg_threshold = CV_BGFG_MOG_BACKGROUND_THRESHOLD; //判断是否为背景点的阈值0.7
params.std_threshold = CV_BGFG_MOG_STD_THRESHOLD;//标准阈值2.5
params.weight_init = CV_BGFG_MOG_WEIGHT_INIT; //权重值0.05
params.variance_init = CV_BGFG_MOG_SIGMA_INIT*CV_BGFG_MOG_SIGMA_INIT; //方差30*30
params.minArea = CV_BGFG_MOG_MINAREA; //？？？
params.n_gauss = CV_BGFG_MOG_NGAUSSIANS; //高斯模型个数
}
else
{
params = *parameters;
}
if( !CV_IS_IMAGE(first_frame) ) //如果第一帧不是图像，则报错
CV_ERROR( CV_StsBadArg, "Invalid or NULL first_frame parameter" );
CV_CALL( bg_model = (CvGaussBGModel*)cvAlloc( sizeof(*bg_model) )); //申请内存空间
memset( bg_model, 0, sizeof(*bg_model) );
bg_model->type = CV_BG_MODEL_MOG; // CV_BG_MODEL_MOG高斯背景模型
bg_model->release = (CvReleaseBGStatModel)icvReleaseGaussianBGModel; //释放内存的函数指针
bg_model->update = (CvUpdateBGStatModel)icvUpdateGaussianBGModel; //更新高斯模型的函数指针
bg_model->params = params;
//申请内存空间
CV_CALL( bg_model->g_point = (CvGaussBGPoint*)cvAlloc(sizeof(CvGaussBGPoint)*
((first_frame->width*first_frame->height) + 256))); //256?
CV_CALL( bg_model->background = cvCreateImage(cvSize(first_frame->width,
first_frame->height), IPL_DEPTH_8U, first_frame->nChannels));
CV_CALL( bg_model->foreground = cvCreateImage(cvSize(first_frame->width,
first_frame->height), IPL_DEPTH_8U, 1));
CV_CALL( bg_model->storage = cvCreateMemStorage());
//初始化
var_init = 2 * params.std_threshold * params.std_threshold; //初始化方差
CV_CALL( bg_model->g_point[0].g_values =
(CvGaussBGValues*)cvAlloc( sizeof(CvGaussBGValues)*params.n_gauss*
(first_frame->width*first_frame->height + 128))); //128?
//n：表示像素点的索引值
//p：表示当前像素对应颜色通道的首地址
// g_point[]：对应像素点、g_values[]：对应高斯模型、variance[]和 mean[]：对应颜色通道
for( i = 0, p = 0, n = 0; i < first_frame->height; i++ ) //行
{
for( j = 0; j < first_frame->width; j++, n++ ) //列
{
bg_model->g_point[n].g_values = bg_model->g_point[0].g_values + n*params.n_gauss;//每个像素点的第一个高斯模型的地址（每个像素对应n_gauss个高斯分布模型）
//初始化第一个高斯分布模型的参数
bg_model->g_point[n].g_values[0].weight = 1; //取较大权重，此处设置为1
bg_model->g_point[n].g_values[0].match_sum = 1;//高斯函数被匹配的次数（？？？）
for( m = 0; m < first_frame->nChannels; m++) //对各颜色通道的方差和均值赋值
{
bg_model->g_point[n].g_values[0].variance[m] = var_init; //初始化较大的方差
bg_model->g_point[n].g_values[0].mean[m] = (unsigned char)first_frame->imageData[p + m]; //赋值为当前像素值
}
//初始化剩下的高斯分布模型的参数
for( k = 1; k < params.n_gauss; k++)
{
bg_model->g_point[n].g_values[k].weight = 0;//各高斯分布取相等且较小权重值，此处取0
bg_model->g_point[n].g_values[k].match_sum = 0;
for( m = 0; m < first_frame->nChannels; m++)
{
bg_model->g_point[n].g_values[k].variance[m] = var_init; //初始化较大的方差
bg_model->g_point[n].g_values[k].mean[m] = 0; //赋值0
}
}
p += first_frame->nChannels;
}
}
bg_model->countFrames = 0;
__END__;
if( cvGetErrStatus() < 0 )
{
CvBGStatModel* base_ptr = (CvBGStatModel*)bg_model;
if( bg_model && bg_model->release )
bg_model->release( &base_ptr );
else
cvFree( &bg_model );
bg_model = 0;
}
return (CvBGStatModel*)bg_model;
}
////////////////////////////////////////////////////////// icvUpdateGaussianBGModel ///////////////////////////////////////////////////////////////
功能：对高斯背景模型变量bg_model进行更新
static int CV_CDECL icvUpdateGaussianBGModel( IplImage* curr_frame, CvGaussBGModel* bg_model )
{
int i, j, k;
int region_count = 0;
CvSeq *first_seq = NULL, *prev_seq = NULL, *seq = NULL;
bg_model->countFrames++;
for( i = 0; i < curr_frame->height; i++ ) //行
{
for( j = 0; j < curr_frame->width; j++ ) //列
{
int match[CV_BGFG_MOG_MAX_NGAUSSIANS];
double sort_key[CV_BGFG_MOG_MAX_NGAUSSIANS];
const int nChannels = curr_frame->nChannels; //通道数目
const int n = i*curr_frame->width+j; //像素索引值
const int p = n*curr_frame->nChannels; //像素点颜色通道的首地址
// A few short cuts
CvGaussBGPoint* g_point = &bg_model->g_point[n];
const CvGaussBGStatModelParams bg_model_params = bg_model->params;
double pixel[4];
int no_match;
for( k = 0; k < nChannels; k++ ) //拷贝各通道颜色分量值
pixel[k] = (uchar)curr_frame->imageData[p+k];
no_match = icvMatchTest( pixel, nChannels, match, g_point, &bg_model_params );
//判断高斯背景模型更新帧数是否达到设置值win_size（？？？）
（初始更新阶段和一般更新阶段在更新处理过程中是不同的，其中定义初始更新阶段为帧数小于win_size）
if( bg_model->countFrames == bg_model->params.win_size ) //一般更新阶段
{
icvUpdateFullWindow( pixel, nChannels, match, g_point, &bg_model->params );
if( no_match == -1)
icvUpdateFullNoMatch( curr_frame, p, match, g_point, &bg_model_params );
}
else
{
icvUpdatePartialWindow( pixel, nChannels, match, g_point, &bg_model_params );
if( no_match == -1)
icvUpdatePartialNoMatch( pixel, nChannels, match, g_point, &bg_model_params );
}
icvGetSortKey( nChannels, sort_key, g_point, &bg_model_params );
icvInsertionSortGaussians( g_point, sort_key, (CvGaussBGStatModelParams *)&bg_model_params );
icvBackgroundTest( nChannels, n, p, match, bg_model );
}
}
//foreground filtering
//filter small regions
cvClearMemStorage(bg_model->storage);
//cvMorphologyEx( bg_model->foreground, bg_model->foreground, 0, 0, CV_MOP_OPEN, 1 );
//cvMorphologyEx( bg_model->foreground, bg_model->foreground, 0, 0, CV_MOP_CLOSE, 1 );
cvFindContours( bg_model->foreground, bg_model->storage, &first_seq, sizeof(CvContour), CV_RETR_LIST );
for( seq = first_seq; seq; seq = seq->h_next )
{
CvContour* cnt = (CvContour*)seq;
if( cnt->rect.width * cnt->rect.height < bg_model->params.minArea )
{
//delete small contour
prev_seq = seq->h_prev;
if( prev_seq )
{
prev_seq->h_next = seq->h_next;
if( seq->h_next ) seq->h_next->h_prev = prev_seq;
}
else
{
first_seq = seq->h_next;
if( seq->h_next ) seq->h_next->h_prev = NULL;
}
}
else
{
region_count++;
}
}
bg_model->foreground_regions = first_seq;
cvZero(bg_model->foreground);
cvDrawContours(bg_model->foreground, first_seq, CV_RGB(0, 0, 255), CV_RGB(0, 0, 255), 10, -1);
return region_count;
}
//////////////////////////////////////////////////////////// icvMatchTest ////////////////////////////////////////////////////////////////
功能：将当前像素与个高斯分布进行匹配判断，如果匹配成功，则返回相应高斯分布的索引值
static int icvMatchTest( double* src_pixel, int nChannels, int* match,
const CvGaussBGPoint* g_point,
const CvGaussBGStatModelParams *bg_model_params )
{
int k;
int matchPosition=-1;
for ( k = 0; k < bg_model_params->n_gauss; k++) match[k]=0; //高斯分布匹配标识数组初始化置0
for ( k = 0; k < bg_model_params->n_gauss; k++)
{
double sum_d2 = 0.0;
double var_threshold = 0.0;
for(int m = 0; m < nChannels; m++) //计算当前高斯分布各通道均值与像素点各通道值相减
{
double d = g_point->g_values[k].mean[m]- src_pixel[m];
sum_d2 += (d*d);
var_threshold += g_point->g_values[k].variance[m];
} //difference < STD_LIMIT*STD_LIMIT or difference**2 < STD_LIMIT*STD_LIMIT*VAR
var_threshold = _model_params->std_threshold*bg_model_params->std_threshold*var_threshold;
//匹配方程为：或者
if(sum_d2 < var_threshold)
{
match[k] = 1; //匹配时标识置1
matchPosition = k; //存储匹配的高斯分布索引值
break; //一旦匹配，就终止与后续高斯分布的匹配
}
}
return matchPosition; //返回匹配上的高斯分布索引值
}
//////////////////////////////////////////////////// icvUpdateFullWindow ////////////////////////////////////////////////////////////
功能：更新各高斯分布的权重值（对于匹配上的高斯分布要增大权值，其余的减小权值），如果存在匹配上的高斯分布，还要更新其均值和方差。
static void icvUpdateFullWindow( double* src_pixel, int nChannels, int* match,
CvGaussBGPoint* g_point,
const CvGaussBGStatModelParams *bg_model_params )
{
const double learning_rate_weight = (1.0/(double)bg_model_params->win_size); //学习率α
for(int k = 0; k < bg_model_params->n_gauss; k++)
{
//若match[k]=0，则权重ω的更新公式：
//若match[k]=0，则权重ω的更新公式：
g_point->g_values[k].weight =
g_point->g_values[k].weight + (learning_rate_weight*((double)match[k] -g_point->g_values[k].weight));
if(match[k]) //更新匹配的高斯分布的参数
{
//参数学习率
double learning_rate_gaussian = (double)match[k]/(g_point->g_values[k].weight*(double)bg_model_params->win_size);
for(int m = 0; m < nChannels; m++)
{
const double tmpDiff = src_pixel[m] - g_point->g_values[k].mean[m];
//均值μ更新公式为：
g_point->g_values[k].mean[m] =
g_point->g_values[k].mean[m] +(learning_rate_gaussian * tmpDiff);
//方差更新公式为：
g_point->g_values[k].variance[m] = g_point->g_values[k].variance[m]+
(learning_rate_gaussian*((tmpDiff*tmpDiff) - g_point->g_values[k].variance[m]));
}
}
}
}
//////////////////////////////////////////////////// icvUpdateFullNoMatch ////////////////////////////////////////////////////////////
功能：当前像素点与所有高斯分布都不匹配时，需要将比值最小的高斯分布替换为新的高斯分布（权值小、方差大），其余的高斯分布保持原来的均值和方差，但权值需要减小。
static void icvUpdateFullNoMatch( IplImage* gm_image, int p, int* match,
CvGaussBGPoint* g_point,
const CvGaussBGStatModelParams *bg_model_params)
{
int k, m;
double alpha;
int match_sum_total = 0;
//new value of last one
g_point->g_values[bg_model_params->n_gauss - 1].match_sum = 1; //将新的高斯分布的match_sum置为1
//get sum of all but last value of match_sum
for( k = 0; k < bg_model_params->n_gauss ; k++ )
match_sum_total += g_point->g_values[k].match_sum;
//设置新的高斯分布的参数
g_point->g_values[bg_model_params->n_gauss - 1].weight = 1./(double)match_sum_total; //给新的高斯分布设置一个较小的权值，即1.0/ match_sum_total
for( m = 0; m < gm_image->nChannels ; m++ )
{
// first pass mean is image value
g_point->g_values[bg_model_params->n_gauss - 1].variance[m] = bg_model_params->variance_init; //初始化一个较大的方差
g_point->g_values[bg_model_params->n_gauss - 1].mean[m] = (unsigned char)gm_image->imageData[p + m]; //将当前像素值作为均值
}
//更新其余高斯分布的参数
alpha = 1.0 - (1.0/bg_model_params->win_size);
for( k = 0; k < bg_model_params->n_gauss - 1; k++ )
{
//更新权值的公式为：
g_point->g_values[k].weight *= alpha;
if( match[k] ) //对于匹配的高斯分布，权值更新公式为
g_point->g_values[k].weight += alpha;
}
}
//////////////////////////////////////////////////// icvUpdatePartialWindow ////////////////////////////////////////////////////////////
功能：更新各高斯分布的权重值（对于匹配上的高斯分布要增大权值，其余的减小权值），如果存在匹配上的高斯分布，还要更新其均值和方差。
static void icvUpdatePartialWindow( double* src_pixel, int nChannels, int* match, CvGaussBGPoint* g_point, const CvGaussBGStatModelParams *bg_model_params )
{
int k, m;
int window_current = 0;
for( k = 0; k < bg_model_params->n_gauss; k++ )
window_current += g_point->g_values[k].match_sum;
for( k = 0; k < bg_model_params->n_gauss; k++ )
{
g_point->g_values[k].match_sum += match[k];
double learning_rate_weight = (1.0/((double)window_current + 1.0)); //increased by one since sum
g_point->g_values[k].weight = g_point->g_values[k].weight +
(learning_rate_weight*((double)match[k] - g_point->g_values[k].weight));
if( g_point->g_values[k].match_sum > 0 && match[k] )
{
double learning_rate_gaussian = (double)match[k]/((double)g_point->g_values[k].match_sum);
for( m = 0; m < nChannels; m++ )
{
const double tmpDiff = src_pixel[m] - g_point->g_values[k].mean[m];
g_point->g_values[k].mean[m] = g_point->g_values[k].mean[m] +
(learning_rate_gaussian*tmpDiff);
g_point->g_values[k].variance[m] = g_point->g_values[k].variance[m]+
(learning_rate_gaussian*((tmpDiff*tmpDiff) - g_point->g_values[k].variance[m]));
}
}
}
}