In the use of opencv BackgroundSubtractorMOG2 gaussian mixture modeling, can detect moving targets, but after long time still a moving target will be to the new detection not to come out in the background, could you tell me how to stationary target detection using modeling

CodePudding user response:

People think this is a little contradictory, background deduction is used to detect movement, stationary object does not have movement in the undetectable, of course,
You have two ideas:
1. Near the motion detection target disappear USES the second secondary test
2. The initial background background subtraction division don't update
Humble opinion, hope to have help!

CodePudding user response:

Upstairs said the thinking of two feasible, if your camera fixed, save the initial of the first frame directly, after the contrast and the first frame to do all time not good

CodePudding user response:

2, on the first floor of the idea is to use background subtraction division, the frame of the first frame is not updated then and after comparison, so that we can detect later into the target of movement and stillness, but the use of model checking, I think is there any way

CodePudding user response:

Moving target moving, keep records of his position, when the target, is using his last position, I don't know at no

CodePudding user response:

The function of the study the frog eyes?

CodePudding user response:

Agree with the fourth floor