SECCHI DISK MEETS UNDERWATER COMPUTER VISION: ADVANCING THEORY AND METHOD

Underwater visibility has long been a subject of interest for sailors, fisherfolk, coastal navigators, and scientists, who, over time, have developed various tools for its measurement and prediction. Among the most enduring is the Secchi disk, a simple white disk lowered into the water using a marked line, whose depth of disappearance serves as a visibility proxy. Although the theory behind Secchi disk visibility (i.e. Secchi disk depth) was formulated in the 1950s and has undergone several revisions since then, the current model still omits important factors regarding the visual perception of the observer, physical characteristics of the disk, and the effect of dynamic environmental conditions. Here, we propose a holistic version of Secchi disk visibility theory that bridges these gaps, leveraging advances in underwater computer vision. We formulate a Secchi disk image formation model that explicitly accounts for the characteristics of the observer’s visual system, as well as the size and pattern of the disk. Then, using physically-based Monte Carlo simulations, we assess the effect of different optical properties of the water body and sea surface conditions on the estimated Secchi disk depth values. In addition, we modernized the disk itself, motorizing it, adding a depth sensor, and replacing the human observer with a camera. Powered by our new image formation model, this modernized Secchi disk device estimates spectral attenuation coefficients from RGB photographs, providing objective and actionable data for scientists, citizen scientists, and coastal communities.  

Presentation Preference: Oral

Primary Presenter: Amir Hadad, University of Haifa (amirhadadad@gmail.com)

Authors:

Amir Hadad, University of Haifa  (amirhadadad@gmail.com)

Arseny Kargin, University of Haifa  (arseny@academ.org)

Derya Akkaynak, University of Haifa  (dakkaynak@univ.haifa.ac.il)