Bathymetry retrieval from CubeSat image sequences with short time lags

The rapid expansion of CubeSat constellations could revolutionize the way inland and nearshore coastal waters are monitored from space. This potential stems ability CubeSats to provide daily imagery with global coverage at meter-scale spatial resolution. In this study, we explore unique opportunity improve retrieval bathymetry offered by CubeSats, specifically those PlanetScope constellation. orbital design constellation enables acquisition image sequences short time lags (from seconds hours). characteristic allows multiple images be captured during a period steady bathymetric conditions, especially in dynamic environments like rivers. We hypothesize that taking ensemble mean sequence can enhance compared standard single-image analysis. Along existing optimal band ratio analysis (OBRA) algorithm, also use new neural network-based depth (NNDR) technique infer both individual time-averaged images. two methodologies evaluated using field data five different river reaches depths up 15 m top-of-atmosphere (TOA) radiance bottom-of-atmosphere (BOA) surface reflectance products. Despite low spectral resolution concerns about radiometric quality imagery, accuracy assessment based on in-situ comparisons indicates (0.52 < R2 0.7 for NNDR method) retrieve ? 10 clear water conditions. proposed averaging consistently improves over single was found outperform OBRA, illustrating importance leveraging all bands through machine learning approaches. TOA provided more robust results than BOA OBRA technique, but minimally impacted type product.