Detection of Moving Objects in Earth Observation Satellite Images: Verification

TL;DR

This paper presents a method to detect and measure the velocity of moving objects in Earth observation satellite images by analyzing multi-spectral data with varying acquisition times, validated using high-altitude balloon imagery.

Contribution

The paper introduces a novel technique to determine relative acquisition times and detect moving objects in satellite images with non-uniform temporal sampling.

Findings

Method successfully measures apparent velocity of objects.

Validated with high-altitude balloon images crossing the US and Colombia.

Enables detection of moving objects in existing satellite archives.

Abstract

In multi-spectral images made by Earth observation satellites that use push-broom scanning, such as those operated by Planet Labs Corp., moving objects can be identified by the appearance of the object at a different locations in each spectral band. The apparent velocity can be measured if the relative acquisition time between images in different spectral bands is known to millisecond accuracy. The images in the Planet Labs archive are mosaics of individual exposures acquired at different times. Thus there is not a unique acquisition time for each spectral band. In an earlier paper, we proposed a method to determine the relative acquisition times from the information in the images themselves. High altitude balloons provide excellent targets to test our proposed method because of their high apparent velocity due to the orbital velocity of the satellite and geometric parallax in images aligned to the level of the ground. We use images of the Chinese balloon that crossed the US in February, 2024 as well as images of an identical balloon over Colombia to test our method. Our proposed method appears to be successful and allows the measurement of the apparent velocity of moving objects from the information available in the archive.