Stratospheric balloon earth observation gathered imagery classification through deep learning

Abstract

Earth observation, also known as remote sensing, is the collection of data about the Earth’s surface and atmosphere using various remote sensing platforms, such as satellites equipped with imaging instruments. The field of computer vision has been increasingly employed for satellite imagery analysis to extract meaningful information from the data collected. However, the cost of launching and maintaining space-based missions can be prohibitive for certain applications, particularly those requiring low-cost testing. An alternative approach that has gained traction in recent years is the use of stratospheric balloons, which are capable of collecting data at high altitudes at a fraction of the cost and time required for space-based missions. This article presents a workflow for implementing a deep learning-based image classification system for stratospheric balloon imagery. In that sense, the proposed system aims to determine the quality of the images captured, with the ultimate goal of utilizing them for science communication and promoting aerospace projects.