**2. Global Appearance Descriptors**

As mentioned in the previous section, global appearance descriptors constitute an interesting alternative for mapping and localization. In this work, the robot moves along the floor plane, and it captures images using a hyperbolic mirror, which is mounted over a camera along the vertical axis. This section details three methods to describe the global appearance of a set of panoramic scenes *IM* <sup>=</sup> {*im*1, *im*2, ..., *imN*} where *imj* <sup>∈</sup> <sup>R</sup>*Mx*×*My* . After using each description method, a descriptor for each image is calculated; thus, the database is composed of a set of descriptors, *D* = { #»*d*1, #»*d*2, ..., # »*dN*} where each descriptor is #»*dj* <sup>∈</sup> <sup>C</sup>*l*×<sup>1</sup> and corresponds to the image *imj*.

The remainder of the section presents the global appearance techniques used throughout the paper and the homomorphic filter, which is used as a pre-treatment for the images.
