Machine Learning Archive - Blog

GeoNode Serves as AI Inference Proxy

June 30, 2025 by Henning Bredel Leave a Comment

Filed Under: 52°North, Machine Learning Tagged With: AI, debris detection, geodata, GeoNode, KI, OpenAPI, PlasticObs+, real-time data, remote sensing

Supporting machine learning on multisensor data from airborne remote sensing

Authors: Henning Bredel and Eike Hinderk Jürrens

The PlasticObs+ Project is part of the BMUV (Federal Ministry for the Environment, Nature Conservation, Nuclear Safety and Consumer Protection) funding initiative “AI lighthouse projects for the environment, climate, nature and resources”. It investigates machine learning methods and tools on multi-sensor data from airborne remote sensing to combat plastic waste in oceans and rivers. The project is part of the funding program “AI-Lighthouses for Environmental Protection” (funding code: 67KI21014A).

more >

Development of an LLM-based Assistant for Searching for Geodata

March 20, 2025 by Simeon Wetzel Leave a Comment

Filed Under: 52°North, Innovation Challenge, Machine Learning Tagged With: large language model, LLM, Metadata, Neural Networks, RDI, SDI

Searching for relevant data is a fundamental task in geospatial and research data infrastructures (SDI/RDI) that demands precise and accurate results. Traditional search systems, which rely primarily on metadata, have notable limitations. These include the need for specialized knowledge and the inability to comprehend the contextual meaning of user queries. As a result, non-expert users often struggle to efficiently find the most relevant datasets. Furthermore, traditional metadata-based searches typically rely on exact keyword matching, which fail to capture the full semantic intent behind a query.

To address these limitations, we will explore a novel approach to geospatial data search using a Large Language Model (LLM) based framework in this blog post.

more >

Hydrological time series forecasting with Neural Networks

April 20, 2022 by Sebastian Drost 2 Comments

Filed Under: 52°North, Deep Learning, Geostatistics Tagged With: Deep Learning, hydrology, Neural Networks, Time series forecasting

Bochum students explore deep learning techniques for rainfall runoff predictions

In recent years, the applicability of Machine Learning approaches for hydrological use cases has reached great attendance. Especially long short-term memory (LSTM) networks show comparable results to conceptual models when being applied to hydrological time series forecasting problems. Now, Interdisciplinary students from Bochum explored innovative deep learning techniques for rainfall-runoff modeling.

more >

Harvesting Ship’s Speed and Environmental Information

May 25, 2021 by Sufian Zaabalawi Leave a Comment

Filed Under: 52°North, Deep Learning, Geostatistics, Machine Learning Tagged With: ais, data analytics, data mining, data scraping, machine learning, ML, speed over ground, vessel traffic, weather routing

The MariData project was recently launched to improve the energy efficiency of ship operations and to reduce emissions. Fuel consumption regarding vessel traffic is affected by many factors, namely, the main and auxiliary engines, the propulsion system, ship hull, propellers, seakeeping performance, as well as weather and sea conditions. In order to tackle these problems, we divide the task of building a weather routing system into subproblems and handle each subproblem separately. Finding the correlations between the speed of the vessel and weather conditions would lead to a better understanding of the weather conditions affecting fuel consumption, therefore, we obtained speed data of the ship’s Automatic Identification Systems (AIS), then retrieved weather and sea information for each data point using timestamp and geographic locations. The combination of these data sources enabled us to build a model to better understand the inherent relationships.

more >