Velocity Analysis of Moving Objects in Earth Observation Satellite Images Using Multi-Spectral Push Broom Scanning

TL;DR

This paper introduces a novel method for estimating the velocities of moving objects in Earth observation satellite images by analyzing multi-spectral data and acquisition timing differences, enabling dynamic object tracking without precise timestamps.

Contribution

It presents a new approach to velocity estimation using multi-spectral push broom satellite data, expanding the utility of commercial satellite imagery for dynamic analysis.

Findings

Successfully estimated aircraft velocities matching onboard transponder data

Demonstrated the method's potential despite data proprietary limitations

Extended satellite imagery applications to dynamic object tracking

Abstract

In this study, we present a method for detecting and analyzing the velocities of moving objects in Earth observation satellite images, specifically using data from Planet Labs' push broom scanning satellites. By exploiting the sequential acquisition of multi-spectral images, we estimate the relative differences in acquisition times between spectral bands. This allows us to determine the velocities of moving objects, such as aircraft, even without precise timestamp information from the image archive. We validate our method by comparing the velocities of aircraft observed in satellite images with those reported by onboard ADS-B transponders. The results demonstrate the potential, despite challenges posed by proprietary data limitations, of a new, useful application of commercial satellite data originally intended as an ongoing, once-daily survey of single images covering the entire land-area of the Earth. Our approach extends the applicability of satellite survey imagery for dynamic object tracking and contributes to the broader use of commercial satellite data in scientific research.