Automatically detect specific changes over large areas  

How to be the first-to-know of changes that occur on specific area ?   

Ask us anything about the problem you want to solve using earth observation, we have a solution. 

​

10 years of huge progress in Computer Vision and Earth Observation helped to achieve great performance to classify automatically changes that are useful for you. All types of images (optical, thermal and radar) can be analysed in order to detect, classify, predict and trigger actions related to your process. A combination of all sensors and time series data allows you to achieve great performance on your use cases.

​

For instance, you could be concerned by compliance issues in terms of security or public health on areas that belong to you and host your industrial operations.

We perform pixel level semantic segmentation tasks on top of pre trained AI models trained with self-supervised learning methodology on satellite images. Self-supervised model captures good knowledge of land cover and the hierarchy between the information, fostering transfer learning to classification tasks with limited amounts of labelled data.  

Machine learning and deep learning can detect but also predict. Based on the new data and strengthened by the patterns learned from historical data you can prevent damages from occurring. Raising alerts when situations in a complex environment are aligned to increase the risk of injuries or damages will not only help you to act directly on the ground but also point out procedures that have to be redesigned. 

The systems we design and implement for our customers learn continuously from the new data and more importantly detect when the precision starts deviating from the results obtained at their release.

​

​