Terrain assessment is a key aspect for autonomous exploration rovers, surrounding environment recognition is required for multiple purposes, such as optimal trajectory planning and autonomous target identification. In this work we present a technique to generate accurate three-dimensional semantic maps for Martian environment. The algorithm uses as input a stereo image acquired by a camera mounted on a rover. Firstly, images are labeled with DeepLabv3+, which is an encoder-decoder Convolutional Neural Networl (CNN). Then, the labels obtained by the semantic segmentation are combined to stereo depth-maps in a Voxel representation. We evaluate our approach on the ESA Katwijk Beach Planetary Rover Dataset.
Evaluation of 3D CNN Semantic Mapping for Rover Navigation
Chiodini, S
;Pertile, M;Debei, S
2020
Abstract
Terrain assessment is a key aspect for autonomous exploration rovers, surrounding environment recognition is required for multiple purposes, such as optimal trajectory planning and autonomous target identification. In this work we present a technique to generate accurate three-dimensional semantic maps for Martian environment. The algorithm uses as input a stereo image acquired by a camera mounted on a rover. Firstly, images are labeled with DeepLabv3+, which is an encoder-decoder Convolutional Neural Networl (CNN). Then, the labels obtained by the semantic segmentation are combined to stereo depth-maps in a Voxel representation. We evaluate our approach on the ESA Katwijk Beach Planetary Rover Dataset.File | Dimensione | Formato | |
---|---|---|---|
2020_Metroaerospace_1.pdf
Accesso riservato
Tipologia:
Published (Publisher's Version of Record)
Licenza:
Accesso privato - non pubblico
Dimensione
15.03 MB
Formato
Adobe PDF
|
15.03 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
2006.09761v1.pdf
accesso aperto
Tipologia:
Preprint (AM - Author's Manuscript - submitted)
Licenza:
Altro
Dimensione
11.56 MB
Formato
Adobe PDF
|
11.56 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.