无法在BackgroundSubtractorMOG2中设置用户参数

时间:2012-10-12 14:17:38

标签: c++ opencv background-subtraction

OpenCV库版本2.42。我想在BackgroundSubtractorMOG2对象中设置一个参数,例如

BackgroundSubtractorMOG2 bgr;  

// the following doesn't work because 'nmixtures', 'backgroundRatio' 
// and 'fVarMin' are a protected members.
bgr.nmixtures = 3;   
bgr.backgroundRatio = 0.9;
bgr.fVarMin = 5; 

// the following works 
bgr.set('nmixtures', 3); 

// both of the following lines will give a run-time error 
// `Access violation reading location 0x0000000000000008.`
bgr.set("backgroundRatio", 0.9);  
bgr.set("fVarMin", 5);     

backgroundRatiofVarMin是控制算法的参数。用户应该能够根据documentation更改这些参数。

如何设置BackgroundSubtractorMOG2

的参数

编辑正如下面的答案中正确提到的,这是OpenCV中的一个错误。该错误已在OpenCV版本2.4.6中修复。

4 个答案:

答案 0 :(得分:9)

我刚刚查看了OpenCV源代码,发现文件/modules/video/src/video_init.cpp中有趣的初始化。这是:

CV_INIT_ALGORITHM(BackgroundSubtractorMOG2, "BackgroundSubtractor.MOG2",
    obj.info()->addParam(obj, "history", obj.history);
    obj.info()->addParam(obj, "nmixtures", obj.nmixtures);
    obj.info()->addParam(obj, "varThreshold", obj.varThreshold);
    obj.info()->addParam(obj, "detectShadows", obj.bShadowDetection));

似乎可以使用方法set设置这四个参数。

还要查看文件modules/video/src/bgfg_gaussmix2.cpp,该文件具有BackgroundSubtractorMOG2类。它包含以下字段:

float fVarInit;
float fVarMax;
float fVarMin;
//initial standard deviation  for the newly generated components.
//It will will influence the speed of adaptation. A good guess should be made.
//A simple way is to estimate the typical standard deviation from the images.
//I used here 10 as a reasonable value

fVarMin(您想要更改)设置为:

fVarMin = defaultVarMin2

在两个构造函数中。以下是所有这些:

static const float defaultVarInit2 = 15.0f; // initial variance for new components
static const float defaultVarMax2 = 5*defaultVarInit2;
static const float defaultVarMin2 = 4.0f;

有趣的是,这个值没有在任何其他文件中使用,因此现在似乎无法改变它。您可以将此问题直接发布到OpenCV bugtracker

答案 1 :(得分:4)

是的,bgr.set("nmixtures",3);应该有效。 BackgroundSubtractorMOG2继承自cv::Algorithm,因此您使用cv:Algorithm::getcv::Algorithm::set来访问这些参数。你尝试过这个并不起作用吗?

答案 2 :(得分:2)

由于这些参数受到保护,派生类可以访问它们。我创建了一个派生类来设置所有必要的参数。

struct BackgroundModel2ParameterBlock {
int nmixtures;
float backgroundRatio;
float varThresholdGen;
float fVarInit;
float fVarMin;
float fVarMax;
BackgroundModel2ParameterBlock(void) :
  nmixtures(3),
  backgroundRatio(0.6),
  varThresholdGen(6.25),
  fVarInit(256),
  fVarMin(256),
  fVarMax(9e2)
{ }
};

class BackgroundModel2 : public cv::BackgroundSubtractorMOG2 {
private:
  BackgroundModel2ParameterBlock m_param;
};

BackgroundModel2::BackgroundModel2(BackgroundModel2ParameterBlock param):
  BackgroundSubtractorMOG2(),
  m_param(param)
{
  nmixtures = m_param.nmixtures;
  backgroundRatio = m_param.backgroundRatio;
  varThresholdGen = m_param.varThresholdGen;
  fVarInit = m_param.fVarInit;
  fVarMin = m_param.fVarMin;
  fVarMax = m_param.fVarMax;
}

答案 3 :(得分:1)

android中,使用算法函数:setDoublesetIntsetBool: 这有效:

mBgMog2 = new BackgroundSubtractorMOG2(mHistory,mMog2Threshold );  
mBgMog2.setInt("nmixtures" , 3);
mBgMog2.setDouble("fVarInit" , 80.0);
mBgMog2.setDouble("fTau" , 0.2);
mBgMog2.setDouble("fVarMin" , 200.0);
mBgMog2.setDouble("fVarMax" , 80.0);
mBgMog2.setBool("detectShadows",false);