Remote Sensing Applications

🧠 GPT changed language. Clay might change the way we understand Earth. Clay is an open-source foundation model for Earth: trained on massive amounts of satellite imagery across location and time. It transforms the complexity of environmental data into powerful embeddings that can be used to: ✅ Identify land cover, crop types, or urban expansion ✅ Detect change like wildfires, floods, or deforestation ✅ Power downstream models for prediction, classification, and mapping ✅ Serve as a backbone for custom geospatial AI pipelines The result? A model that understands Earth the way LLMs understand language. Training models is tough, plus you need access to massive amounts of data. As foundational models start to get better, the data backbone being built by Cloud-Native Geospatial Forum (CNG) data and computing systems that can leverage these models like those we are working on at Wherobots can help bring these models to global scale. This is bigger than just another geospatial model. It’s a signal that foundation models are coming to remote sensing, and with them, a new paradigm: 🧠 Pre-trained models that can be adapted everywhere 📡 Build models with fewer labels 🌱 Tackle climate, agriculture, and environmental challenges with speed If you’re working in geospatial AI, Earth observation, or climate data: Clay is worth watching. And using. It's open source and live on Hugging Face and GitHub. The geospatial foundation model era is bound to be an exciting one. 🌎 I'm Matt and I talk about modern GIS, geospatial data engineering, and how spatial thinking is changing. 📬 Want more like this? Join 5k+ others learning from my newsletter → forrest.nyc

𝐆𝐞𝐨𝐬𝐩𝐚𝐭𝐢𝐚𝐥 𝐌𝐚𝐜𝐡𝐢𝐧𝐞 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐀𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬 Machine learning is no longer just an analytical tool - it’s becoming the backbone of geospatial intelligence. From wildfire prediction and groundwater mapping to disease forecasting, carbon estimation, and urban sprawl detection, these papers show how spatial data + ML are reshaping environmental risk assessment, urban analytics, agriculture, and climate science. If you're working at the intersection of GIS, remote sensing, and AI - this collection of recent research papers is worth bookmarking. My tutorials: https://lnkd.in/dXpjUz3K ⬇️ Full list in the end ⬇️ 1. Exploration of Geo-Spatial Data and Machine Learning Algorithms for Robust Wildfire Occurrence Prediction https://lnkd.in/dtTW_iau 2. Enhancement of Groundwater Resources Quality Prediction Using an Improved DRASTIC Method and Machine Learning https://lnkd.in/dNhTsieN 3. Remote Sensing-Based Forest Cover Classification Using Machine Learning https://lnkd.in/dZfAUZs4 4. Forest Age Estimation Based on a Machine Learning Pipeline Using Sentinel-2 and Auxiliary Data https://lnkd.in/dr3c79-P 5. Factors of Acute Respiratory Infection Among Under-Five Children Using Machine Learning Approaches https://lnkd.in/d6DUxbAh 6. SAR Image Integration for Multi-Temporal Wetland Dynamics Analysis Using Machine Learning https://lnkd.in/dY4--gep 7. Effects of Non-Landslide Sampling Strategies in Landslide Susceptibility Mapping https://lnkd.in/d7mRkFWv 8. Enhancing Co-Seismic Landslide Susceptibility and Risk Analysis Through Machine Learning https://lnkd.in/dtsigmG8 9. 10-m Scale Chemical Industrial Parks Map Along the Yangtze River Based on Machine Learning https://lnkd.in/dS9qGi68 10. Geospatial Distribution and Machine Learning Algorithms for Assessing Surface Water Quality in Morocco https://lnkd.in/dwxSamAt ... 20. Wheat Crop Genotype Identification Using Multispectral Radiometer Data and Machine Learning 21. Geospatial Data for Peer-to-Peer Communication Among Autonomous Vehicles Using Optimized ML Algorithms ... 𝐅𝐮𝐥𝐥 𝐥𝐢𝐬𝐭: https://lnkd.in/dNC9VA7E

The world is changing. 2024 was the first year to surpass 1.5 degrees Celsius. Climate change, deforestation, pollution—the challenges aren’t new. We have been hearing about them for years. But can AI become a true game-changer in addressing them? In 2024, natural disasters caused $368 billion in economic losses worldwide, with 60% of these damages uninsured. Despite this, AI-powered tools are beginning to shift how we respond. ➡️ AI-powered tools, like Google Earth’s Cloud Score+, are stepping up to fill critical gaps. By providing clearer images of ecosystems obscured by clouds, such innovations make monitoring the environment faster and more accurate. ➡️ AI Algorithms now track polar ice melt, analyze deforestation trends, and even alert authorities to illegal logging within hours. ➡️  In Brazil, AI-driven deforestation monitoring cut illegal activities by 20% last year, saving millions of hectares of rainforest. These advancements highlight how AI turns raw satellite data into tools for immediate action. ➡️  Researchers are deploying AI-powered drones to track marine species, improving conservation efforts. Smart fishing systems, driven by AI, help reduce bycatch by distinguishing between target fish and other marine life. ➡️ Air quality monitoring is being transformed by AI. Google’s Air View+ system in India has improved air quality in cities like Aurangabad by 50% over three years, proving how AI can drive cleaner urban environments. The possibilities are limitless, from personalized climate action plans to autonomous drones monitoring remote ecosystems. But technology alone isn't enough. AI gives us the tools to combat environmental crises, but the question remains: how will you contribute? Whether adopting eco-friendly habits, supporting AI initiatives, or staying informed, every action counts. What do you think? #AI #climatechange #technology

🚀 AlphaEarth Foundations (AEF) - New from Google DeepMind I keep looking out for interesting usecases of AI. Deepmind folks are at it again. 📄 Paper: AlphaEarth Foundations on arXiv (https://lnkd.in/giHUwe2d) --- 🌍 What is AlphaEarth Foundations? AEF is a foundation model for Earth observation that turns sparse and messy satellite, climate, LiDAR, and even text data into dense embeddings at 10 m² resolution. These embeddings provide a universal feature space for mapping and monitoring the planet, outperforming all previous approaches — reducing mapping errors by ~24% on average. And the best part? The embeddings are already available as annual global datasets (2017–2024) for free: 👉 Earth Engine Data Catalog: Google Satellite Embedding V1 Annual - https://lnkd.in/g6dcv4-M --- 🛠 Why does this matter? (weekend project ?) For places like Bengaluru, India (or any fast-changing city), AEF makes it possible to: - Track urban growth and land use change with very few ground samples. - Monitor lakes and wetlands for encroachment and seasonal changes. - Map flood risk by combining rainfall, elevation, and land cover. - Identify urban heat islands and vegetation loss. - Support peri-urban agriculture with low-shot crop type classification. - Study biodiversity shifts (tree species, invasive plants) by linking with GBIF/iNaturalist data. In short, it’s like having a plug-and-play geospatial backbone — ready to support everything from city planning to climate adaptation. --- 🔧 For the Geeks Want to try it out? You can get started in minutes using Earth Engine + Python: 📘 Earth Engine Python Quickstart Docs - https://lnkd.in/g9zBBPJv 🌐 This is a big step toward planetary-scale AI for environmental monitoring — making high-quality maps possible even when labels are scarce. --- Further reading : 1. https://lnkd.in/gsXU2BqS 2. https://lnkd.in/gxJpqS6b --- Authors: Christopher Brown, Michal Kazmierski, Valerie Pasquarella, William J. Rucklidge, Masha Samsikova, Chenhui Zhang, Evan Shelhamer, Estefania Lahera, Olivia Wiles, Simon Ilyushchenko, Noel Gorelick, Lihui Lydia Zhang, Sophia Alj, Emily Schechter, Sean Askay, Oliver Guinan, Rebecca Moore, Alexis Boukouvalas, Pushmeet Kohli.

Unlocking the Power of GeoAI: From Raw Geospatial Data to Actionable Insights GeoAI is fundamentally changing the way we work with geospatial data. Today, artificial intelligence is not just a research topic, but a practical tool that helps us turn massive amounts of aerial imagery and lidar data into real, actionable information. By combining neural networks with proven photogrammetry and rule-based quality assurance, we can now extract detailed land cover maps, analyze urban surfaces, and even simulate urban climate with a level of precision that was unthinkable just a few years ago. One of the most exciting aspects is how GeoAI enables us to move beyond traditional mapping. With AI-powered segmentation, we can distinguish even the smallest features in urban environments and keep our data up to date. Thanks to TrueOrthos and advanced photogrammetric workflows, geometric distortions are a thing of the past, so data from different times and sensors can be perfectly aligned. This is essential for reliable change detection and multi-source analysis. But the possibilities go even further. Automated analysis of sealed and unsealed surfaces helps cities identify where to prioritize “desealing” for climate resilience. Parcel indexing allows us to aggregate key indicators like green space, building area, or solar installations at any scale, supporting truly data-driven decisions in urban planning and environmental monitoring. And with urban climate simulation, we can combine pixel-precise land cover data with 3D voxel models and CFD to visualize the effects of new trees, green roofs, or lighter pavements, before any construction begins. Even lidar point cloud classification benefits from GeoAI. By combining AI with rule-based checks and external data sources, we achieve robust, scalable, and quality-assured 3D mapping, reducing manual effort and increasing reliability, even in complex or changing environments. GeoAI is already a productive, scalable approach that is shaping the sustainable, data-driven development of our cities and landscapes. With annual updates and hybrid workflows, we ensure that results are not only precise and up to date, but also trusted and actionable. If you want to learn how to turn your geospatial data into valuable information using GeoAI, just reach out or send me a message. Let’s move from data to information, using GeoAI. 💡 Comment | Like | Share   👉 Follow me (Dr. Uwe Bacher) for more Information on exciting topics from the world of geospatial #GeoAI #Geospatial #AerialImagery #Lidar #UrbanPlanning #AI #SmartCities

#QualityMonday – When a Millimeter Turns into a Megarisk During inline production monitoring last week, our team identified a defect that is easy to miss — and expensive to ignore: In multiple modules, the busbar in the positive junction box was not fully covered by AB pottant. Upon inspection, the conductor tab was found to be lifted upward, leaving part of the metal exposed to air instead of fully resting in the pottant. Scope of risk: - 318 modules isolated on the spot - Root cause traced to abnormal lay-up welding and loose clamp mechanisms that caused some busbars to be 4 mm longer and improperly seated 🧨 Why this is a serious reliability concern: When the pottant doesn’t fully encapsulate the lead, it exposes the busbar to: - Moisture ingress - Corrosion and oxidation over time - Increased contact resistance - Potential thermal events near the diode area - Premature junction box failure in the field - Potential safety hazard under moisture And importantly: Once the module is sealed and shipped, this defect is almost impossible to detect without destructive inspection. ✅ Immediate actions taken: - Junction box soldering machine stopped - Welding head position adjusted by +0.5 mm - Operators instructed to flatten any lifted busbars before potting - Entire curing room batch quarantined (318 pcs) Factory added: - Hourly checks on busbar height - Manual flattening at framing - Monthly clamp mechanism maintenance - QC patrol inspections 🔎 The takeaway: A single lifted tab inside the junction box — invisible from the outside — can compromise long-term reliability and safety. Without independent QA oversight, this becomes a field failure waiting to happen. Clean Energy Associates (CEA) once again: ✔️ Detected the defect early ✔️ Prevented 318 risky modules from shipping ✔️ Forced corrective action and 8D root cause resolution Hidden defects are only hidden when nobody is looking. #PVQuality #SolarManufacturing #ReliabilityRisk #QualityAssurance #RootCauseAnalysis #ManufacturingExcellence

Canopy Height Estimation with GeoAI and Pretrained Models Learn how to estimate canopy height from aerial and satellite imagery using pretrained deep learning models with the GeoAI Python package. This step-by-step tutorial shows you how to use Meta's High-Resolution Canopy Height Model with minimal code! Video Tutorial: youtu.be/vsIRTM98qaU Notebook Example: https://lnkd.in/ei26z7fq Meta's Canopy Height Model: https://lnkd.in/e99mFQzq #GeoAI #RemoteSensing #CanopyHeight #DeepLearning #Python #OpenSource

Solar Performance Monitoring: Practical Examples with Fault Analysis To understand how data analysis helps in fault detection and performance optimization, let’s look at real-world scenarios with sample values. Example 1: Underperformance Due to Soiling Losses 🔹 Expected Power Output: 500 kW 🔹 Actual Power Output: 450 kW 🔹 Performance Ratio (PR) = (450 / 500) × 100 = 90% ✅ (Good) After a week: 🔹 Expected Power Output: 500 kW 🔹 Actual Power Output: 400 kW 🔹 PR = (400 / 500) × 100 = 80% ⚠ (Declining) 🔹 Soiling Loss Estimate: 10-12% 📌 Diagnosis: Increased dust accumulation on panels is reducing efficiency. 📌 Action: Schedule panel cleaning and monitor PR improvement. Example 2: Inverter Failure Leading to Downtime 🔹 Total Plant Capacity: 1 MW 🔹 Number of Inverters: 10 (Each handling 100 kW) 🔹 Before Issue: • Expected Output: 950 kW (considering minor losses) • Actual Output: 940 kW ✅ (Good Performance) 🔹 After Issue: • Expected Output: 950 kW • Actual Output: 840 kW ⚠ (Significant Drop) • Inverter Logs: • Inverter 6: No output • Fault Code: Overvoltage error 📌 Diagnosis: One inverter failure resulted in a 100 kW generation loss. 📌 Action: Restart the inverter remotely via SCADA, if unsuccessful, perform on-site inspection for hardware issues. Example 3: Faulty Solar Panel String Detection 🔹 Total Plant Capacity: 500 kW 🔹 Number of Strings: 50 (Each handling 10 kW) 🔹 Normal Operation: • Each string generating 9.5 - 10 kW 🔹 Current Readings: • 49 Strings: 9.8 kW ✅ (Normal) • 1 String: 6.5 kW ⚠ (Underperforming) 📌 Diagnosis: Possible issues include: ✅ Loose connection in the junction box. ✅ Module degradation in one or more panels. ✅ Partial shading from nearby object. 📌 Action: Perform IR thermographic scanning to check for hotspots and replace faulty panels if needed. Example 4: Impact of High Temperature on Efficiency 🔹 Ambient Temperature: 45°C 🔹 Panel Temperature: 70°C 🔹 Power Output Drop: 5-6% compared to normal conditions 📌 Diagnosis: High temperatures reduce panel efficiency due to the negative temperature coefficient (-0.5% per °C above 25°C). 📌 Action: ✅ Install cooling solutions (e.g., water mist or ventilation). ✅ Use bifacial or high-temperature-resistant panels for future installations. Example 5: Grid Instability Causing Shutdown 🔹 Normal Grid Voltage: 415V 🔹 Recorded Grid Voltage: 470V ⚠ (Overvoltage) 🔹 Inverter Logs: “Grid Overvoltage Protection Activated – Shutdown Initiated” 📌 Diagnosis: ✅ Overvoltage from the grid triggered the inverter’s protective shutdown. ✅ Possible transformer tap setting issue or reactive power injection problem. 📌 Action: ✅ Coordinate with the grid operator to stabilize voltage fluctuations. ✅ Enable reactive power control in the inverter to manage voltage spikes. #SolarMonitoring #DataAnalytics #IoT #SCADA #PredictiveMaintenance #RenewableEnergy #IliosPower

🌟 ADVANCED INSPECTION & TESTING METHODS – HOW DO THEY COMPARE? 🌟 At 2DegreesKelvin, we’re on a mission to promote the importance of advanced inspection & testing methods. Our latest resource, the Advanced Inspection Methods Overview, is here to help you understand how these techniques compare. 🚀☀️ The table breaks down inspection methods into: ✅ Currently Possible 🤔 Questionable / Possible in the Future 🚫 Not Possible (Yet!) 🚁 Although 2DK are massive supporters & advocates of site-wide Thermography surveys, and believe these should be standard practice for commissioning and operational phases of every solar farm. The technology has limits in terms of forensic level diagnosis and root-cause determination, only being able to identify, inverters, strings, modules & diodes out and then ‘hot spots’. There’s some great evidence now that it can indicate PID through the checker board patterning of hot spots, but this isn’t conclusive and in some cases can be other electrical/system conditions. 🩻 Electroluminescence (EL) on the other hand, is by far, the most advanced and valuable inspection & test method available to us in the solar market. It can identify over 90% of all known defects, damages and deterioration phenomena. That’s not to say it’s a ‘Silver Bullet’, its not. Typically, an accompanying Flash test in a calibrated mobile lab, will super-charge your understanding of the modules health, and is extremely complementary to the EL images, and enables us to have the fullest understanding of what’s going on with a module. 🆕👴 This goes for brand new modules, as well as operational ones. Particularly if a checking the Chinese OEM has provided quality, your considering a warranty claim, or your faced with an insurable loss event, then the EL & Flash combination, almost certainly, will provide the optimal depth and detail of the issue to enable key decisions to be made going forward. ❌ What's not in the table is 'High-Pot' testing. This is a function that we have on our test machines which injects a current through the module, we attached 4 x probes to modules frame. We are then able to measure leak-current or (Insulation Resistance - RISO) in dry and wet conditions. This is making an impact on the operational fleet right now, to provide clear evidence of RISO failing modules for warranty claims. 2DK do recommend a combination of these, to be used in different circumstances. So this table is designed for people in the large-scale solar sector, who need to know more about the health, performance and safety of their modules. 💡 By using these advanced techniques, you can: ✔️ Gain deeper insights into your solar assets ✔️ Minimise investment risks ✔️ Maximise system performance 💬 What do you think? Which inspection methods have you seen the most success with? 🔗 If you would like to talk to our expert team, please get in touch: info@2degreeskelvin.org #makesolarbetter #mission70

Thermal #Drones + #AI don’t just inspect solar farms — they reveal invisible power loss. Manual checks = slow, reactive, expensive. #Thermal + #AI + #Geospatial #Intelligence = fast, autonomous, and measurable. Imagine spotting a single faulty solar panel in a 100-acre farm— --- in minutes, not days. --- with exact geo-coordinates. --- and estimated power loss. 1. Identify Radiometric thermal cameras (e.g. DJI Mavic 3T / DJI Matrice 350 RTK + H20T) capture solar farms during solar noon to detect thermal anomalies. 2. Detect Deep learning models (YOLO, U-Net, Transformer encoders) analyze thermal signatures to classify fault types and predict severity levels, including:  • Hotspots  • PID  • String failures  • Soiling & shading  • Bypass diode faults Thermal anomalies are correlated with I-V curve behavior → energy yield estimation → real $ impact. 3. Locate Each fault is geo-referenced to its exact panel row and column → generating actionable work orders for field teams instead of vague reports. 4. Typical Faults & Losses ------------------------------------------- • Defect              --------> Power Loss   ------------------------------------------- • Hotspots          ---------->  5–15 %        • PID               ---------->     10–30 %       • Bypass Diode Failure ------>  15–25 %      • Soiling / Shading  ---------->    5–20 %     • String Failure     ---------->    30–100 %   -------------------------------------------- Why it matters: ✅ 70 % faster inspections ✅ Predictive energy loss modeling ✅ Fault-to-panel traceability ✅ Lower downtime & increased ROI #AI + #Thermal #Drones are redefining solar O&M — from detection to diagnosis to dollars. The complete solution is available on AeroMegh Intelligence- designed and developed by us!