Advances in wireless networking technologies are enabling networks of wireless camera networks, which have applications ranging from surveillance to situational awareness for emergency responders, to tele-immersion and tele-surgery. Though wireless camera networks are a natural extension of the idea of a sensor network, camera networks introduce numerous challenges because of the high-dimensionality of images. These challenges include efficient transmission of multiple images of the same scene as well as the unsupervised analysis and aggregation of an ensemble of images. In this talk I will introduce tools based on a combination of multiple view geometry and harmonic analysis to address some of the issues in a camera network, wireless or otherwise. Whereas multiple view geometry is traditionally based on discrete geometric features, we demonstrate that the space of constraints which encode multiple view relations lie in a space amenable to a type of Fourier transform. This framework enables fast convolution algorithms with which we can analyze camera geometries in a global, non-parametric manner. No familiarity with harmonic analysis or computer vision is assumed.