Upstream Tech的动态

Every tree, every forest, every quarter. We’re thrilled to release Forest Carbon Monitoring, an AI-powered dataset that measures changes to forest carbon and canopy across the globe at 3 meter resolution every quarter! This dataset presents a precise, scalable, and affordable solution for measuring forest dynamics and is an indispensable tool for emerging voluntary carbon markets and mitigating deforestation. Forest Carbon Monitoring leverages Planet’s extensive archive of PlanetScope imagery, a global library of airborne and spaceborne LiDAR data, and AI processing to equip stakeholders with a cost-effective way to monitor forested areas—scaling from a single tree to the entirety of the Amazon Rainforest. Read more: https://go.planet.com/fcm

The world’s forests are vital repositories of #biodiversity, home to millions of #Indigenous peoples, and critical #carbon sinks. If we want to protect them, we have to be able to monitor them with scientific rigor, in real time, at a resolution that allows all stakeholders to take effective action. That’s why yesterday was such a big day for the world’s forests, as Planet unveiled our global, high-resolution forest monitoring products, which use #satellite imagery, remote sensing data and #artificialintelligence to scientifically measure the carbon in every tree on Earth, down to a resolution of just a few meters -- and illuminate how forest carbon is changing. This breakthrough capability has the potential to transform #carbon markets and reporting, by providing scientifically accurate, trusted, consistent, scaled measurements for every forest carbon credit and every REDD+ project. Not only can we detect deforestation and degradation, but every time it occurs, we can assess how much carbon has been lost to the system. (And conversely, when we engage is #afforestation or #reforestation, we can see how much has been gained or restored.) The same data can be used to anticipate and reduce #climate-linked #disaster risks like #wildfires, and strengthen the #conservation and governance the world’s #forests. (These forest monitoring datasets will be produced quarterly, for the whole planet Earth.) This is yet another example of Planet’s ongoing journey to produce what we call “Planetary Variables,”?— spatially-explicit inventories of the world’s #carbon, #biomass, #biodiversity, #water, #population, the #humanfootprint — which are the building blocks with which can both measure and fulfull the ambitions of the #ParisAgreement, the #SDGs and the #GlobalBiodiversityFramework, #sustainable #supplychains, and much else. You can get started with this dataset today. Planet partners like BeZero Carbon and Upstream Tech are already building amazing things with it! You can learn more details in the post below! Congratulations for this milestone go especially to David Marvin, Christopher Anderson, Kyle Gertridge, Tara O'Shea, and the larger Planet team that made this possible. A transparent world is being born before our eyes!

A driving use case when Planet was first founded was to count every single tree on the Planet. We've been routinely providing the data to do that for some time now (see NICFI - Norway's International Climate and Forest Initiative), but this new release allows us to quantify the carbon in those trees at a level of detail never done at this scale. Incentives drive everything and data like this is needed to properly account for our natural capital in the global economic system.

Very cool way of mapping out urban canopies and green spaces.

#14 FoMo-Bench Forest monitoring is one of the most important tasks for Earth’s ecosystems. Deep learning can help towards this goal. ? For this purpose, this paper proposes a Forest Monitoring Benchmark (FoMo-Bench), made of 15 diverse datasets encompassing satellite, aerial, and inventory data, covering a variety of geographical regions, and including multispectral, RGB, SAR, and LiDAR data with various temporal, spatial and spectral resolutions. ? The authors also propose FoMo-Net, a baseline masked image modeling foundation model. ? what I liked most: the important task a note: the comparisons do not include other remote sensing pre-trained models paper: https://lnkd.in/dDbTFhw5

?? An AI bot made a summary of key highlights of PureForest, a massive multimodal dataset for tree species classification that I made publicly available last month: "- Largest and Most Comprehensive Lidar Dataset:?Encompassing 449 distinct monospecific forests and 18 different tree species, it’s designed for both 3D and 2D data analytical approaches. - Significant Contributions for Ecological Studies:?By providing high-quality data, PureForest enables the creation of accurate forest maps, facilitating better conservation strategies." The robot got it right! The AI-generated summary ?? https://lnkd.in/e9gQX2TE The data paper on ArXiv ?? https://lnkd.in/ezvBXEJJ

Welp. Looks like we can't make anything cool with Meta's tree canopy model. The sauce is Meta proprietary - but an overview of the GEDI correction step is in section 3.4 of the paper: https://lnkd.in/e9fXbkBw I dare not venture into trying to recreate the correction step it based on the description, but the wizard who is able to will probably be able to make a whooole lot of ??????. Aerial imagery and lidar are available for most of the United States, and we could do inference on that... but once we pull lidar, what is the point of a tree height model if we are telling it how high the trees are? Time to put this model down and look for something else cool to play with. Hats off to Meta for releasing the global dataset on Earth Engine for scientists to use. Hat trick broken? ??

In the wake of increasing climate challenges, assessing forest health has become more critical than ever. ?? At Flai, we are leveraging LiDAR technology and cutting-edge Artificial Intelligence to transform how we detect and assess forest damage, paving the way for more efficient forest management and conservation efforts. By harnessing the power of LiDAR point cloud data, we can automatically identify and classify damaged areas, providing precise insights that are crucial for timely interventions. Our solution enables faster, more accurate, and scalable forest damage assessments, which can support various industries, from environmental agencies to utility vegetation management. ?? Read more about how AI and LiDAR are revolutionizing forest damage assessment here: https://lnkd.in/djf4N5_n #AI #LiDAR #ForestManagement #ConservationTech #DigitalTwins #Sustainability #Innovation #Flai #Geospatial #ForestAssessment

The use of UAV's and remote sensing in mapping complex habitats has been gaining a lot of traction recently. Being able to map species distributions of coastal macroalgae provides critical insights into the health of these coastal ecosystems, as well as identifying and monitoring changes in composition and diversity due to anthropogenic and natural events. We have been developing some novel techniques to identify and map species distributions, abundances, and functions from shores to whole coastline scales using spectral mapping and AI techniques - to enable environmental managers to do just that!

NUVIEW's Acquisition of Astr?a: Advancing Earth Observation with AI NUVIEW acquires Astr?a to integrate AI-driven analytics into its LiDAR satellite constellation, facilitating global distribution of Earth observation data. By leveraging AI, machine learning, and cloud computing, NUVIEW aims to provide centimeter-scale accuracy in geospatial intelligence for defense, climate initiatives, and more. This strategic move enables NUVIEW to diversify across markets, empowering partners with actionable insights for climate resiliency and disaster mitigation. Together, they pioneer a digital expressway for Earth observation data, addressing critical global challenges. NUVIEW's space-based LiDAR, coupled with Astr?a's AI platform, offers scalable analytics for businesses and governments, advancing sustainable development and climate initiatives worldwide. Clint Graumann #NUVIEW #Astr?a #AI #EarthObservation #LiDAR #SatelliteConstellation #GeospatialIntelligence #MachineLearning #CloudComputing #Defense #ClimateInitiatives #GlobalChallenges #ClimateResiliency #DisasterMitigation #SustainableDevelopment #SpaceTechnology #DigitalExpressway