Measure an object's position/orientation/size with single camera on robot arm

I’m studying a vision project that involves positioning an object with an unknown size. The environment consists of a robot arm equipped with a single camera. I have calculated camera’s intrinsic parameters, distortion coefficients, and hand-eye matrix.

To tackle this problem, I plan to utilize the ORB (Oriented FAST and Rotated BRIEF) feature detection algorithm to extract keypoints from the images and determine the 2D position of the object within the pictures. However, I’m unsure about the process of using different robot poses to solve this problem. I’m wondering if OpenCV provides any APIs or functions that can assist me in this regard.


so… what exactly are you having issues with?

sounds like you want to use the robot to move the camera, so you can get various views of the scene.

I want to get the position/orientation/size of object.