English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent studies have shown the benefits of using additional elevation data(e.g., DSM) for enhancing the performance of the semantic segmentation ofaerial images. However, previous methods mostly adopt 3D elevation informationas additional inputs. While in many real-world applications, one does not havethe corresponding DSM information at hand and the spatial resolution ofacquired DSM images usually do not match the aerial images. To alleviate thisdata constraint and also take advantage of 3D elevation information, in thispaper, we introduce a geometry-aware segmentation model that achieves accuratesemantic labeling of aerial images via joint height estimation. Instead ofusing a single-stream encoder-decoder network for semantic labeling, we designa separate decoder branch to predict the height map and use the DSM images asside supervision to train this newly designed decoder branch. In this way, ourmodel does not require DSM as model input and still benefits from the helpful3D geometric information during training. Moreover, we develop a newgeometry-aware convolution module that fuses the 3D geometric features from theheight decoder branch and the 2D contextual features from the semanticsegmentation branch. The fused feature embeddings can produce geometry-awaresegmentation maps with enhanced performance. Our model is trained with DSMimages as side supervision, while in the inference stage, it does not requireDSM data and directly predicts the semantic labels in an end-to-end fashion.Experiments on ISPRS Vaihingen and Potsdam datasets demonstrate theeffectiveness of the proposed method for the semantic segmentation of aerialimages. The proposed model achieves remarkable performance on both datasetswithout using any hand-crafted features or post-processing.

Geometry-Aware Segmentation of Remote Sensing Images via Implicit Height Estimation