LauraNSotomayor🤠🦇
@lauransotomayor
Followers
1K
Following
6K
Media
239
Statuses
4K
🇲🇽|Ex-Athlete Hurdler🇪🇦|MITS Research🇦🇺|∆PhD goal_Geomatic Engineering @Sciences_UTAS~Ecosystem monitoring with #UAS #remotesensing with #AI #deeplearning
Hobart, Tasmania
Joined May 2017
PhD paper is out🤠We advance methods for mapping fractional vegetation cover using a CNN-based U-net on cm-scale UAS RGB & multispec imagery, capturing patterns in sparse, complex vegetation & enabling detailed, scalable mapping in semi-arid ecosystems(1/8)
link.springer.com
Landscape Ecology - Monitoring fractional vegetation cover (FVC) is crucial for assessing ecosystem health and sustainably managing semi-arid rangelands. Field-based methods are resource-intensive,...
8
5
13
RT @robmarkcole: 🎙️ New Podcast & YouTube: TorchGeo 1.0 with Adam Stewart. In this episode, Adam Stewart shares updates on TorchGeo, the wi….
0
6
0
RT @wkentaro_: Confidence is built from keeping promises to yourself, overcoming the seemingly impossible, and seeing steady progress.
0
1
0
RT @LabelmeAI: .@lauransotomayor applied ConvNet and semantic segmentation for land-health monitoring in Australia🇦🇺. Cool application of c….
0
1
0
RT @robmarkcole: Mapping fractional vegetation cover in UAS RGB and multispectral imagery in semi-arid Australian ecosystems using CNN-base….
0
1
0
If you are interested to play with the CNN-based workflow for mapping ecosystems, is available for RGB, multispectral #drone data, but also for more bands/channels such as hyperspectral inputs if required 👇(8/8).
github.com
Mapping Fractional Vegetation Cover (FVC) components by introducing a CNN-based deep learning approach for UAS imagery - LNSOTOM/fvc_composition
1
0
3
Thanks to all the co-authors involved @TerraLuma, Darren Turner, Megan Lewis and @TejaKattenborn Feeling proud of this mission🚀🤓 (7/8).
0
0
2
Implementation of the Monte Carlo Dropout as Uncertainty estimation in FVC segmentation mapping 🗺️#DeepLearning #RemoteSensing.(6/8)
0
0
1
Cm-scale FVC maps🌱from our best CNN models predictions on 3072×3072 px UAS multispectral tiles for each site (site-specific models). (5/8)
0
0
1
Model generalisation & transferability 🛰️ We tested 2 setups (A & B) to see when site-specific models beat generic ones. Expanded discussion on limitations + showed how data augmentation can boost robustness as a proof-of-concept. (4/8)
0
0
1
U-Net models were trained & validated with spatial block CV, including both site-specific and generic models across all sites, to produce more realistic and transferable evaluations. Random CV inflated performance estimates by up to 28%. (3/8)
0
0
2
We tested our CNN approach at 3 AusPlots sites in Calperum Station, South Australia, each representing a different National Vegetation Information System (NVIS) vegetation type: A (low), C (medium), E (dense). Together they cover 29 ha of semi-arid landscape.🌱📍.(2/8)
0
0
4
RT @TejaKattenborn: New publication: Schiefer "Large-scale remote sensing reveals that tree mortality in Germany appears to be greater than….
academic.oup.com
Abstract. Global warming poses a major threat to forests and events of increased tree mortality are observed globally. Studying tree mortality often relies
0
4
0
RT @RS_Haynes: Few months ago but. New paper out! ⚠️🗞️. Here we used an optical flow algorithm to co-register drone-based hyperspectral i….
mdpi.com
Remote sensing from unoccupied aerial systems (UASs) has witnessed exponential growth. The increasing use of imaging spectroscopy sensors and RGB cameras on UAS platforms demands accurate, cross-co...
0
3
0
RT @Wkolby: Happy to share a new review in #LandscapeEcology, "Applications of unoccupied aerial systems (UAS) in landscape ecology . ," e….
0
7
0
0
53
0