PGS’ Post

🛰️☁️🛰️ One of the activities in the AgriDataValue project was to investigate how filtering cloud shadows improves the quality of the input data. Cloud shadows can be a major nuisance when processing #satellite data, as they are more difficult to detect than the clouds that cast them. We approached this problem by projecting #cloud masks onto the #Earth's surface using an estimated average cloud height in the #Sentinel2 L1C scene. For validation, we used the #CESBIO active learning reference data and even compared the results with the #Sen2Cor Scene Classification map. 👀📄 Read our latest blog post - 𝐏𝐫𝐨𝐣𝐞𝐜𝐭𝐢𝐧𝐠 𝐂𝐥𝐨𝐮𝐝 𝐒𝐡𝐚𝐝𝐨𝐰𝐬 𝐟𝐫𝐨𝐦 𝐂𝐥𝐨𝐮𝐝𝐬 - for more details! ➡️ https://lnkd.in/dk7Fk-7k

definitely cloud shadow algorithm helps us in calculations...

DROP: Dust-GS: Dense Point Clouds Matter 🤘 Instead of relying solely on SfM, Dust-GS introduces an innovative point cloud initialization technique that remains effective even with sparse input data. Paper: https://lnkd.in/encWthKY #gaussiansplatting Gaussian Splatting #3DGS #radiancefields

Are sidecar models in service meshes becoming outdated? K8s specialist Alfred Tommy doesn't think so, but there were tons of talks and hype at #KubeConEurope2024 about Istio's sidecar-less alternative: Ambient Mesh. One knock on sidecars is their resource consumption. Since there is a dedicated sidecar proxy container for each application, resource consumption is multiplied across all application instances in the cluster (100 applications --> 100 sidecar proxies). Ambient Mesh addresses this by using a node-level proxy approach, with a single proxy agent running as a DaemonSet on each node to handle traffic for all applications (100 applications --> potentially a single proxy) Of course, sidecar-less has its drawbacks, which Alfred addressed as well when he joined us on Cloud Masters to cover his highlights from KubeCon Europe 2024 in Paris. Check out the full recap here: https://hubs.li/Q02szW_z0 #kubernetes

Global Mapper Pro can thin point clouds, classify, and extract features. Global Mapper Pro can be used seamlessly with aerial and terrestrial point cloud data. Learn about improving the positional accuracy of point clouds in this Industry Showcase: https://ow.ly/SgGC50QyKAL

Photogrammetry is an incredibly useful tool. Everyone carries a camera these days. What no one appreciates is the data that camera is capable of producing. If you understand the basics and apply a little technique, you can produce dense point clouds. Good data doesn't have to be expensive. The cloud below was produced using Capturing Reality. (Yes, that's a dense point cloud NOT a textured mesh).

What is this spectral thing? How do I make it work with all these really cool dyes? Can I really do a 20+ color panel? Yes you can. Check out this Spectral Hotspot Matrix and see how they all fit together. :)

Got my confirmation for yesterday at 12 UTC, though my orientation of the commas was a bit different and probably should have been corrected, because the strongest #convection is usually in a comma’s tail – in most cases. More about the analysis maps like the one below in the near future, I hope. “A #Comma is an active, #mesoscale #cloud spiral in the cold air behind #frontal cloud bands, often with significant weather.” #WeatherAnalysis #SynopticChart

I have conducted plane estimation on point cloud data of a chair using #Open3D. The parts estimated as a plane are displayed in yellow, while the other parts are shown in red. This plane estimation method was presented at #AIMTG under the title “Preprocessing of 3D Point Clouds Using RANSAC”. Slide https://lnkd.in/gbv9UgXm #3DData #PointCloud #RANSAC

🌐 There is huge global demand for data centres driven by cloud computing, high performance computing, #AI and 5G, but the industry faces an increasing number of hurdles. Marianne Toghill, Hein Tonnaer, Angela Kearns, Rob Donell, Adam Waite, Irina Steinberg, Sam Comer, Mike Niekoop and Lotte de Bruin recently headed to Data Centre World and The Tech Show to get insights on what's ahead. 👉 Read our key takeaways: https://okt.to/9JdBIV #DataCentres #TMT