为什么在跟踪线程poseoptimization函数的鱼眼相机模式中添加e->mMovingProbability = pMP->GetMovingProbability()？

[zlthaha]

在Optimizer.cc的943行，poseoptimization函数中，为何在SLAM with respect a rigid body中添加e->mMovingProbability = pMP->GetMovingProbability()，双目和单目为什么没添加：
//Conventional SLAM
if (!pFrame->mpCamera2) {
// Monocular observation
if (pFrame->mvuRight[i] < 0) {
........
........
........
} else // Stereo observation
{
........
........
........
}
}
//SLAM with respect a rigid body
else {
........
........
........

                    e->mMovingProbability = pMP->GetMovingProbability();
                       ........
                    ........
                    ........
                } else { //Right camera observation
                       ........
                    ........
                    ........
                }
            }