How Drones and AI help Hunze en Aa’s to win the Dutch battle against water.

When roughly 60% of your country’s population lies below sea level, and 70% of your economic output is generated there, then there is talk of significant risk. Add to this the spectre of global warming and increasingly extreme weather, and you have the pre-requisites for a perfect storm. There is zero tolerance for failure – Holland has to be sure of its defences. The Netherlands is home to over 3,700 kilometres of primary flood defence systems, with more than 10,000 kilometres of regional ones. Without this complex network of these dykes, flood basins, and sea defences, Holland would be literally be up the creek without a paddle. At any given day, over 400 professional inspectors therefore manually examine these systems, in search for erosion, subterranean voiding, and areas that may allow water to slip through. In the context of global warming and national safety, a more intelligent and effective approached is now called for.

Creating a 3D model of the dykes.

We now equip an industrial drone with the latest LiDAR technology, in combination with an advanced ground-penetrating radar. With LiDAR it is now possible to create a digital 3D model of the dyke, that even distinguishes between different forms of vegetation and rock. As LiDAR scans are very detailed, the tiniest surface cracks and holes can be observed. Then there is the ground-penetrating radar, which uncovers what is going on underneath the surface. These combined forces discover subtle details that can otherwise easily be overlooked with the naked eye. And since areas of concern are precisely marked on the 3D model, it provides a better understanding of both the in and outside condition of the dyke.

AI-driven dyke examinations.

One of the major drawbacks is the time it takes to manually examine these 3D models, whereas another concerns the massive amounts of data that LiDAR produces. To cope with these challenges, Dronepoint has partnered up with HAL24K: one of the world’s leading enterprises in data intelligence solutions. HAL24K delivers one of the highest standards of processing power to date, so that data-hungry liDAR scans can be processed within the hour. And with advanced AI and machine learning systems, potential damages are automatically recognised. This means that dyke inspectors can focus their expertise where it matters most: on the right time, on the right location.

Anticipating instead of reacting.

Processed 3D models are now stored and organised by location in asset-viewer. Over time, it will now be possible to track the development of the asset, by comparing 3D models to one another. And with the help of HAL24K, machine learning systems will open-up new avenues for predictive maintenance, so that damages can be identified very early stage.