Seaborne 3D Cloud Sensing Using Photography from Free-Drifting Floats

Oceans and large lakes typically have meager atmospheric sensing, despite covering much of Earth. This leads to a major meteorological sensing deficiency, particularly under clouds. Satellite observations suffer from occlusion of the cloud base and content beneath. There are experimental land-based systems for three-dimensional (3D) cloud sensing, but yet none over water. We thus introduces a concept to reduce the deficiency: a floating, wide-baseline, multi-view wireless camera network. The overlap of the camera's view fields enables stereoscopic 3D mapping. We confront marine-engineering challenges, such as wave-induced motion blur and rolling shutter distortion. This is achieved by mechanical design of the floats, on-board sensors and algorithms. We pay attention to the communication bottleneck of sensors deployed far at sea, by performing data reduction and 3D analysis onboard the floating system. A twofloat network prototype was deployed in the Mediterranean. It operated autonomously and successfully measured cloud heights. The results were compared to standard meteorological data.