# Compute Homography using Rotation, translation and Camera Matrix

**URL:** <https://forum.opencv.org/t/compute-homography-using-rotation-translation-and-camera-matrix/17078>\
**Category:** Python\
**Tags:** calib3d\
**Created:** [March 30, 2024, 2:28am UTC](https://forum.opencv.org/t/compute-homography-using-rotation-translation-and-camera-matrix/17078 "2024-03-30T02:28:17Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![aRk](https://avatars.discourse-cdn.com/v4/letter/a/94ad74/32.png) [@aRk](https://forum.opencv.org/u/aRk)\
**Post date:** [March 30, 2024, 2:28am UTC](https://forum.opencv.org/t/compute-homography-using-rotation-translation-and-camera-matrix/17078/1 "2024-03-30T02:28:17Z")

</div>

Hello everyone,

I have calibrated two cameras and computed their rotation matrices (R) and translation vectors (t) using solvePnP, based on the corners of a checkerboard pattern observed in the images and the corresponding 3D world positions of these corners, expressed in meters. Now, I wish to compute the homography matrix utilizing R and t, rather than deriving it from point correspondences as findHomography does, even though findHomography gives perfect results for my application. I’m looking for a method to use rotation and translation directly for this purpose.

Additionally, accurately mapping positions from the image to 3D world coordinates could also serve my needs.

I would appreciate any suggestions on how to achieve this.

Thanks