Automated estimate of fish abundance through the autonomous imaging device GUARD1

Video Image Enhancement and Machine Learning Pipeline for Underwater Animal Detection and Classification at Cabled Observatories

“…An image processing pipeline ( Figure 3) was designed and developed based on computer vision tools for enhancing the image contrast and for segmenting relevant image subregions [19,33]. To speed up this process, the images were resized from 3456 × 5184 pixels to a quarter of their size, i.e., 964 × 1296 pixels.…”

“…A binary thresholding value, which was chosen by testing different values, was performed to obtain the mask of the elements in the image [19,33,[40][41][42][43] and different morphological transformations such as closing, opening, and dilation were applied to remove noise.…”

Exo-Ocean Exploration with Deep-Sea Sensor and Platform Technologies

“…High-definition (HD) imaging is widely used in ecological exploration of Earth's deep-sea, and current tools may be used to identify the presence of fauna with sessile or motile morphological characteristics on icy moons, although that possibility is to date still highly uncertain (Newman, 2018). Within this context, fast-developing deep-sea imaging technologies centered on HD photogrammetry, stereo, hyperspectral, miniaturized cameras and low-light vision are established tools that permit assessment of the presence and activity of organisms (e.g., Kokubun et al, 2013;Bicknell et al, 2016;Corgnati et al, 2016, Marini et al, 2018a. These imaging assets could be adapted for the identification of exo-oceanic fauna in a broad range of sizes (i.e., equivalent to our prokaryotes, including bacterial mat formations, as well as micro-eukaryotes, micro-and meso-zooplankton, up to larger multicellular organisms).…”

“…based computer vision for the detection of pelagic organisms (Marini et al, 2018a(Marini et al, , 2018b. Those devices are being used to study a still evanescent life component of our oceans, which is represented by large aggregates (deepscattering layers) of bathymetrically displacing organisms, being hence of utility to scan large volumes of Enceladus' exo-ocean for similar purposes.…”

“…This problem happens also in deep-sea research, where solutions are provided by data science, pattern analysis, and artificial intelligence methodologies (Skiena, 2017;Aguzzi et al, 2019). Simple computer vision algorithms can be executed on board platforms' imaging asset, to identify any subject different from the water or seabed itself (Corgnati et al, 2016;Marini et al, 2018a). General approaches based on image enhancement, differencing, and background subtraction methods can be used to discard irrelevant information (Moeslund, 2012;Peters, 2017); for example, in the case of water column, ice shell, or seabed surfaces, changes in patterns would be slower with respect to traveling objects.…”