SharkTrack: an accurate, generalisable software for streamlining shark and ray underwater video analysis

Abstract

Elasmobranchs (shark sand rays) are critical components of coral reef ecosystems and are often considered indicators of reef health (Roff et al., 2016). Yet, they are experiencing global population declines and effective monitoring of these populations is essential to their protection. Underwater stationary videos, such as those from Baited Remote Underwater Video Stations (BRUVS), are critical for understanding elasmobranch spatial ecology and abundance. However, processing these videos requires time-consuming manual analysis. To address these challenges, we developed SharkTrack, a semi-automatic underwater video analysis software.SharkTrack uses Convolutional Neural Networks and Multi-Object Tracking to automatically detect and track elasmobranchs and provides an annotation pipeline to manually classify elasmobranch species and compute MaxN, the standard metric of relative abundance. We tested SharkTrack on BRUVS footage collected from threecoral reef locations unseen by the model during training, to demonstrate the model’sadaptability and effectiveness in different reef environments. SharkTrack computedMaxN with 89% accuracy over 207 hours of footage. The semi-automatic SharkTrackpipeline required two minutes of manual classification per hour of video, a 95%reduction of manual analysis time compared to traditional methods, estimated conservatively at 42 minutes per hour of video. These results suggest thatSharkTrack can be utilised to monitor elasmobranch populations across diverse coral reef ecosystems. Furthermore, the software’s flexible pipeline could serve as a blueprint for the development of species classifiers beyond elasmobranchs, enabling more comprehensive monitoring of coral reef biodiversity. We provide public access to SharkTrack, aiming to support future research in coral reef and marine conservation.