1: \begin{abstract}

2:             Feature-based homography estimation approaches rely on extensive image processing for feature extraction and matching, and do not adequately account for the information provided by the image.

3: 			Therefore, developing efficient direct techniques to extract the homography from images is essential.

4: 			This paper presents a novel nonlinear direct homography observer that exploits the Lie group structure of $\SL(3)$ and its action on the space of image maps.

5: 			Theoretical analysis demonstrates local asymptotic convergence of the observer.

6: 			The observer design is also extended for partial measurements of velocity under the assumption that the unknown component is constant or slowly time-varying.

7: 			Finally, simulation results demonstrate the performance of the proposed solutions on real images.

8: 	%\lipsum[1]

9: \end{abstract}

10: