| void Impala::Core::Array::Trait::FuncKalman::UpdateKalman | ( | ) | [inline] |
this function does reestimation of the sigmas using the residual
Definition at line 176 of file FunctorKalman.h.
References ComputeResiduals(), InitialiseSigma(), mBeta, mGamma, mInitialised, mOcclusion, mOutliers, mPixelCount, mResidualAvrg, mResidualThisFrame, mSigmaG, mSigmaI, and mSigmaW.
00177 { 00178 //eq.14 00179 mResidualThisFrame /= mPixelCount; 00180 mResidualThisFrame *= mBeta; 00181 if(mResidualThisFrame == 0) 00182 mResidualThisFrame = 1; 00183 00184 if(mPixelCount * mGamma > mOutliers) // if no occlusion 00185 { 00186 mOcclusion = false; 00187 ComputeResiduals(); 00188 if(!mInitialised) 00189 InitialiseSigma(); 00190 //eq.11 00191 mSigmaG = (mSigmaG*mSigmaI)/(mSigmaG+mSigmaI); 00192 //eq.15 00193 mSigmaW = mResidualAvrg - mSigmaI - mSigmaG; 00194 // following found in Hieus code, makes some sense, but only because the 00195 // assumption about residual and sigmas are incorrect (I can't define them, sorry) 00196 if(mSigmaW<0) 00197 mSigmaW = 0; 00198 //eq.9 (prediction of sigmaG for the next time step) 00199 mSigmaG += mSigmaW; 00200 } 00201 else 00202 { 00203 if(!mOcclusion) 00204 { 00205 //undo prediction of sigmaG (eq.9) (only if the last time sigmaG was predicted) 00206 mSigmaG -= mSigmaW; 00207 } 00208 mOcclusion = true; 00209 } 00210 mPixelCount = 0; 00211 mOutliers = 0; 00212 }
Here is the call graph for this function:
