This communication is about a simplifed method to build image mosaics from aerial images. Getting very exact, orthonormally projected mosaics (orthomosaics) from aerial imagery is nowadays a standard technique performed with the so called “Photogrammetry” [1,7] applications. Nevertheless, photogrammetry applications are normally expensive proprietary software (despite some open-source options of less quality and difcult to use), creating orthomosaics is normally a slow process. In this paper, we try to develop an easy method able to create a simplifed mosaic useful for some applications (mainly computer vision applications where very exact geometry measurements are not necessary). Furthermore, our method is much faster than classic photogrammetry as it is based on algorithms used to synthesize panoramic images [2]: control points are automatically derived using SURF [3,4] algorithm and then used to compute geometric transformations [5] so that warped images coincide in the overlapping part. Images are warped and overlapped to form a whole mosaic.