sadshark's profile - activity

Hello everyone, I would like to know if an effective method to find an image in a video already exists. Let me explain:

I have a video quite linear, where a camera is moving along an axis in a room. I also have a picture of the same room (somewhere on the axis, and slightly different) and I would like to determine which frame in the video matches the best this picture.

I dont know if a method already exists for that, Otherwise i may have ideas but i am open to suggestions. thank you!

Actually cx and cy are always the coordinates of the center of my image

Hello everyone, I am using a stereo camera to process the images in order to work with point clouds. I am using a method proper to the PCL library to create those point clouds (shame on me) but anyway, here is my problem: The camera sensor resolution is 1/3.2" 752x480, it has a baseline of 12cm and a focal length of 2.17 mm.

I managed to convert the focal length to pixels and pass it as well as the other parameters to the function that I am using to create the point clouds. It could be faster if I was reducing the initial image resolution: I would like to pass from 752x480 to 263x168 (35% of the initial resolution) in order to process the clouds faster. I am currently stucked on this problem: How to transform the previous parameters to adapt them to this new resolution. I checked the OpenCV camera calibration documentation but I didnt find what i was looking for (maybe I didnt looked into it properly).

could you help me on this?

thank you in advance

edit: I dont know if my problem has a solution but I am kind of desperate