API src

Found 2 results.

Hydrobiid mudsnail image dataset: Raw stereomicroscope, processed, and segmented images of abundant brackish snails from a mesocosm experiment with multiple temperature treatments

We examined the growth response of a brackish snail (Hydrobiidae) against multiple temperature treatments in a mesocosm located beside the Alfred Wegener Institute Wadden Sea Station on Sylt (55°01′19.2″N, 8°26′17.7″E). Bulk sediments were collected south of Pellworm (54° 31' 55.83"N, 8° 42' 40.36"E) at low tide on March 22, 2022, transferred to mesh-lined crates and introduced to mesocosm tanks (170 cm × 85 cm × 1800 L). Experimental warming treatments were conducted using three heaters per tank (Titanium heater 500 W, Aqua Medic, Bissendorf, Germany). The full specifications for the mesocosms are already published (Pansch et al., 2016). Throughout the experimental warming period, four sampling events (March 30, April 25, May 24, June 20) were conducted to core sediments. Sediment cores were washed and sieved (1mm mesh size) to disaggregate infauna. Individuals were separated for the common hydrobiid mudsnail, which were subsequently imaged in groups on a typical petri plate under stereomicroscopy. A semi-automatic object segmentation and size measurement approach was developed to rapidly differentiate and measure individuals from images. Segmentation was highly accurate and precise against manual length measurements (end-to-end; mm) collected in ImageJ for 4595 snails. Scaling the segmentation method across the full dataset estimated >40,000 snails and presented a complex species-specific response to warming. The enclosed dataset represents all raw, processed, and segmented images (n= 3201) produced by this study.

Physikalische Modelle in der wissensbasierten Analyse von Fernerkundungsbildern (FSP Bildverarbeitung und Mustererkennung)

Im Rahmen des Forschungsschwerpunktes 'Theory and Applications of Image Processing and Pattern Recognition' des FWF werden wissensbasierte Bildanalysemethoden auf die Interpretation von Fernerkundungsbildern angewendet. Es werden physikalische Modelle der Aufnahme von Fernerkundungsbildern (unter Betonung der radiometrischen Aspekte) formuliert und Bildanalysemethoden entwickelt, die auf der Inversion dieser Modelle beruhen. Teilaspekte dieses Projekts betreffen insbesondere auch Methoden der Subpixelanalyse.

1