Semantic segmentation in forested environments: 1 student

The purpose of this study is to develop a robust semantic scene understanding system of unstructured environments. The proposed system should address several inherent natural factors such as shadows, illumination, glare. Expected segmentation output: vegetation, sky, water, sand and mud.

Dataset: adapt and develop datasets (e.g. Freiburg forest dataset)

Contact: Ahmedt Aristizabal, David (Data61, Black Mountain)

Updated:  10 August 2021/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing