Ocean Ledger reposted this
Today, thanks to the explosive growth in #EarthObservation of the last two decades, we have access to a multitude of public and commercial satellites that can help us identify and map coastal ecosystems like #seagrasses. More recently, this growth has led to the development and broad deployment of lightweight #drones, creating a highly-flexible and diverse remote sensing toolkit. Choosing your remote sensing instrument can be challenging - often requiring a mixture of domain and technical expertise, but there are a number of parameters you can use to assess what the right choice or combination is: 🛰️ Spatial resolution (i.e. size of the pixel) 🛰️ Temporal resolution (i.e. frequency of acquisition) 🛰️ Spectral characteristics & resolution (i.e. number of bands and their spectral width across sensor's wavelengths) 🔘Geographical scale (i.e. area of interest) 💰 Cost of image(s) (e.g. available or tasked) The effect of varying spatial resolution of various optical remote sensing instruments (e.g. public and commercial satellites, and a very high-resolution drone), along with their pros and cons, are showcased in the image below, over a shallow seagrass meadow and sandy seabed in Lesvos, Greece. #Coastal ecosystems are inherently complex in terms of their environment, physiology and #ecology, and, most often, more than one phenomenology (i.e. modality) is required to map, measure and monitor them, across space and time, at high accuracy, inference and granularity. In Ocean Ledger, we are resolving these complexities by combining multi-modal (i.e. optical, radar, lidar) data-agnostic (in source, platform and resolution) analytics with AI, cloud computing, modelling and field data, both for onshore and underwater coastal ecosystems. Leveraging the full spectrum of #remotesensing technologies allows us to optimize our proprietary modular workflows and provide robust, standardized downstream insights.
Nice overview! I would add an additional bullet point to your list of criteria: data quality. Even if it has the exact temporal, spatial and spectral needs for you, there's very limited value to a dataset if the data has too much noise to be useful! Characterisation of sensor performance extremely important for some applications.
This is a great summary and a cool graphic that captures the challenges of balancing marketing claims with what you actually need. With Drones, I would add fixed-wing aircraft flying with Airborne LiDAR Bathymetry, then you can negate the small ground area coverage limitation and improve the uncertainty of the optical algorithms by including measured depth values and an active light source (the laser). Of course, it can be logistically challenging and it will cost money! What value do we place on natural capital?
I love it!!!
Co-Founder & CSO at Ocean Ledger
1wFigure produced by Levi Westerveld/GRID-Arendal (2019)