The powerful combination of AI and Earth Observation Data

The powerful combination of AI and Earth Observation Data:

01 February 2019

At DHI GRAS we routinely conduct coffee talks, where our experts present their latest research and development within advanced image analytics, data science or other trends within Earth Observation.

This week our AI expert Philip Graae demonstrated results of applying Deep Learning on Earth Observation Data.

By using principles of transfer learning, we have trained a state-of-the-art object detection neural network to detect and segment small and very detailed objects within earth observation image data.

With proper tuning and by using techniques such as data augmentation, it is possible to get good results with just a small amount of training data as input to the AI.

We rely on such deep learning techniques when traditional image analytics is not sufficient to extract the information from the imagery.

Contact Philip Graae or gras@dhigroup.com for more information on how our AI developments and skillsets could benefit your needs.

January 25, 2019February 6, 2019

The bi-annual Danish Marine Research event

The Danish Marine Research event:

25 January 2019

DHI GRAS presenting state-of-the-art marine and coastal research

Today is the last day of the bi-annual Danish Marine Research event, covering state-of-the-art research within marine and coastal topics.

In the session ‘New Technology in Marine Research and Monitoring’ we presented our new high-resolution Danish bathymetry model providing a first ever full national cover of shallow water bathymetry in Denmark and the status of our 2-year activity on high-resolution mapping of submerged aquatic vegetation at national scale funded by the VELUX FONDEN.

It is clear that the acceptance of satellite-based information as a valuable tool for monitoring the marine and coastal zone is increasing and we look forward to further work on increasing the acceptance through continued delivery of high-quality services and data products.

Contact us at gras@dhigroup.com if you want to know more about how we use satellite data to monitor the coastal zone or read about the project here.

January 9, 2019June 14, 2019

Mapping aquatic vegetation in Denmark

Mapping aquatic vegetation in Denmark:

09 January 2019

For the first time, the aquatic environment around the Danish coastline will be mapped using satellite images.

By using satellites, the task is both easier to carry out and gives a better spatial overview.

This will help to strengthen the protection of our coastal water environment.

The project has been made possible thanks to a grant from the VELUX Foundation.

Watch the video below to learn more!

December 14, 2018June 14, 2019

Open-Source Software Development

Open-Source Software Development:

14 december 2018

There are countless of open source products we use and love at DHI GRAS, and it is very dear to us to give back to the community.

Therefore, we are thrilled to announce that we have just released our new, light-weight tile server Terracotta as free software under an MIT License!

Terracotta directly serves your (cloud-optimized) GeoTIFF files, so you don't need to apply any pre-processing to get your raster data online, and it supports serverless deployment on AWS Lambda, so you can serve your data without any permanent infrastructure.

We urge you to check it out, and look forward to your feedback.

December 13, 2018May 21, 2019

AI for Earth

AI for Earth:

13 december 2018

To get more yield with fewer resources, precision is key – especially with a precious resource like water. One way to know how much water crops really need is to measure how much water evaporates from the soil and plant surfaces. This is known as the evapotranspiration (or ET) rate, which helps farmers improve their watering efficiency

Torsten Bondo

Torsten Bondo: Business development manager and senior remote sensing engineer at DHI GRAS in Denmark.

His team aims to use machine learning and satellites to support irrigation development and improve crop water efficiency in Uganda together with the Ugandan geo-information company Geo Gecko. The goal is to contribute to food security, poverty alleviation and economic growth.

The project that been chosen as one of the eleven changemakers to receive the Microsoft and National Geographic AIforEarth Innovation Grant.

December 5, 2018June 14, 2019

Satellites and offshore wind farms

Satellites and offshore wind farms?:

05 december 2018

A match made in space..

With an increasing number of satellites launched in recent years and advances in the image analysis methods, remote sensing data has proven itself as a cheap and very valuable source of information for planning, construction and monitoring of offshore wind farms (OWF).

Mapping of water depths and seabed dynamics

Understanding of the seabed dynamics is highly relevant when planning to reduce potential risks of exposure of cables and infrastructure. Using advanced image analytics, it is possible to convert raw satellite images into quantified information of the seabed around OWFs. The information can either be derived as absolute water depth values in shallow waters (down to 25m), or as relative information highlighting the geomorphological dynamics, such as estimation of migration rates (metres per year) of sand waves.

DHI has successfully delivered this type of information for OWFs along the US east coast, in the North Sea and in Taiwan Strait.

Detailed monitoring of maintenance activities in near-real-time

As most satellites are in constant orbit around the Earth, they provide a mean for getting situational awareness when needed. Optical and radar (SAR) satellites can provide very detailed imagery down to 25 cm in detail level, and with no HSE risks associated. The images are a highly cost-efficient way of getting updated intel on specific OWF sites, allowing monitoring of specific maintenance or similar activities. The images can be delivered within just hours after the acquisition and thus provide the situational overview in near-real-time.

With the increasing amount of satellites, it is now possible to get a daily snapshot in very high resolution of any OWF site in the world. This capacity will be even further increased within the next few years where sub-daily (down to hourly) imagery is expected.

More accurate and reliable metocean data

For years, satellite remote sensing data has provided incredible value to metocean processes and characteristics in relation to offshore wind. With the use of altimeters and radar sensors mounted on satellites, it is possible to derive several different parameters related to quantification of metocean conditions, such as winds, waves and sea surface height – today this is even possible very close to a coastline.

Not only is satellite data of significant value for wind resource estimation, it also allows for the establishment of an accurate and robust environmental design basis for building offshore wind turbines.

Today, it is common to apply aggregated measurements from different types of satellites and multiple passes in calibration and validation of numerical hydrodynamics and wave models, the so-called hind cast models – in combination with in situ synoptic observations. The outcome of the hind cast models are accurate time series (typically 30-40 years) of metocean data to be used for planning, design, construction and operation of an offshore wind farm.

Such data is readily available on modern metocean data platforms such as DHI’s MetOcean Data Portal, where long term data can be downloaded easily and quickly for subsequent processing and analysis. Before data (e.g. wave data) is downloaded, the data is always instantly validated against satellite data to assure the quality and robustness.