We are pleased to announce the launch of our latest research initiative focusing on automated horizon line detection. Accurate horizon estimation is a fundamental challenge in computer vision, playing a critical role in autonomous navigation, attitude estimation, and image stabilization for both aerial and maritime platforms. Students are developing robust algorithms designed to perform reliably under diverse environmental conditions. You can find an initial demonstration of our progress in the video below, which showcases our preliminary detection results in real-time scenarios.

You can view preliminary results from recordings made on the streets of Budapest by clicking here. Horizon lines are estimated via vanishing points.

Video created by Márton Koós.

Explore More

New Open-Source Release: AffineExtractor Toolbox

We are pleased to announce the public release of AffineExtractor, a new Python-based computer vision toolbox developed by our research group. The repository is now available on GitHub for the

Visit of Professor Weisi Lin / NTU

On 13 May, Professor Weisi Lin from Nanyang Technological University, Singapore, visited our research group. He gave a seminar entitled ‘Exploring Opportunities with 3D Point Clouds‘. Abstract3D point clouds (PCs)

ICRA 2026: New Paper on Real-Time Surface Normal Estimation

We are thrilled to announce that our latest research, “Robust and Real-time Surface Normal Estimation from Stereo Disparities using Affine Transformations”, was successfully presented at the prestigious ICRA 2026 conference