I'm pretty sure that there's no software that'll create camera movement data that can be imported into Poser. The only way you can probably do it, is by opening the video footage as a background in Poser. Then you'll have to animate Poser's camera to match how the camera moved in the video footage. I suppose if you used Poser's 'ground plane' with the 'vanishing point' (line?) option turned on, it won't be impossible. (At the very least, you need to get the gound plane matched perfectly with the video.) The point is to build a temporary virtual scene that matches the scene in the video. take the objects out when you're ready to render.

Probably a LOT of work, tho. With all the camera's key frames being set to 'linear' in Poser, you'll probably still find yourself having to keyframe the camera's position for almost EVERY frame of the video. It depends on how much the camera was moved when the video was being made.

Use some of the 'shape props' (cubes; spheres) to match the position of stationary objects in the video, as closely as possible. I take it that you don't have any measurements from the video camera to any objects that are in the video that was made. (?) You're at a serious disadvantage, right from the start. Remember that for the next time.

You can save 'keyframe data files' from inside Poser's keyframe editor. DO THIS, CONSTANTLY! Otherwise, if your scene file goes corrupt, you'll be starting over with NOTHING.

As you get the ground plane and Poser objects (name them!) matched up with the video, write down the positions and dimensions you see under Poser's dials for them. If the scene file goes bad, you'll still have that, at least.

So you have a point of reference, to know what you're working with...
There's a program called "MatchMover" that does what you want to do. All it creates, is a camera position and rotation data file for programs like 'Max', 'Lightwave', etc..
It costs about $7,500.00