Discover
What's

NEW

FIND WHAT YOU'RE LOOKING FOR

Resource Type

Industry

Application

Arrival Cities: Classifying Urban Poor Areas with Satellite Imagery

Hannes Taubenböck and his team at DLR are using Very High Resolution (VHR) satellite imagery, supplied by European Space Imaging, combined with auxiliary surveys to develop a base model classification system for the shape and structure of urban poor areas around the world. Slum, favela, tenement…whatever you call it, nearly

Automated Beach Litter Detection Using Satellite Imagery and Machine Learning

The Tama Group demonstrates how they use machine learning to detect trash along beaches from VHR satellite imagery and generate density maps to aid in cleanup efforts. Traditionally it has been a human user who analyzes satellite imagery for changes, but this process is slow and time-consuming. Whether detecting trash

AW3D Enhanced

The ability to analyse the elevation of terrain and surface features in high resolution is a crucial component of most geospatial applications. European Space Imaging has partnered with NTT Data to offer AW3D Enhanced elevation models. By leveraging the world’s highest resolution satellite imagery from the Maxar WorldView Constellation, AW3D

AW3D Standard

European Space Imaging has partnered with NTT Data to offer AW3D Standard – a global elevation model derived from millions of satellite images acquired by the Advanced Land Observing Satellite from JAXA. Featuring high geolocational accuracy without any Ground Control Points (GCP) and resolutions of 2.5 m or 5 m,

Scroll to Top

Architecture of ResNet34-UNet model

UNet architecture for semantic segmentation with ResNet34 as encoder or feature extraction part. ResNet34 is used as an encoder or feature extractor in the contracting path and the corresponding symmetric expanding path predicts the dense segmentation output.

Architecture of VGG16-UNet model

UNet architecture for semantic segmentation with VGG16 as the encoder or feature extractor. VGG16 is used as an encoder or feature extractor in the contracting path and the corresponding symmetric expanding path predicts the dense segmentation output.

Architecture of ResNet34-FCN model

In this model, ResNet34 is used for feature extraction and the FCN operation remains as is. The feature of ResNet architecture is exploited where just like VGG, as the number of filters double, the feature map size gets halved. This gives a similarity to VGG and ResNet architecture while supporting deeper architecture and addressing the issue of vanishing gradients while also being faster. The fully connected layer at the output of ResNet34 is not used and instead converted to fully convolutional layer by means of 1×1 convolution.

Architecture of VGG16-FCN model

In this model, VGG16 is used for feature extraction which also performs the function of an encoder. The fully connected layer of the VGG16 is not used and instead converted to fully convolutional layer by means of 1×1 convolution.

X