Development and application of a video-mosaic survey technology to document the status of coral reef communities

Abstract

The recent decline in the condition of coral reef communities worldwide has fueled the need to develop innovative assessment tools to document coral abundance and distribution rapidly and effectively. While most monitoring programs rely primarily on data collected in situ by trained divers, digital photographs and video are used increasingly to extract ecological indicators, provide a permanent visual record of reef condition, and reduce the time that divers spend underwater. In this study, we describe the development and application of a video-based reef survey methodology based on an algorithm for image registration and the estimation of image motion and camera trajectory. This technology was used to construct two-dimensional, spatially accurate, high-resolution mosaics of the reef benthos at a scale of up to 400 m(2). The mosaics were analyzed to estimate the size and percent cover of reef organisms and these ecological indicators of reef condition were compared to similar measurements collected by divers to evaluate the potential of the mosaics as monitoring tools. The ecological indicators collected by trained divers compared favorably with those measured directly from the video mosaics. Five out of the eight categories chosen (hard corals, octocorals, Palythoa, algal turf, and sand) showed no significant differences in percent cover based on survey method. Moreover, no significant differences based on survey method were found in the size of coral colonies. Lastly, the capability to extract the same reef location from mosaics collected at different times proved to be an important tool for documenting change in coral abundance as the removal of even small colonies (<10 cm in diameter) was easily documented. The two-dimensional video mosaics constructed in this study can provide repeatable, accurate measurements on the reef-plot scale that can complement measurements on the colony-scale made by divers and surveys conducted at regional scales using remote sensing tools.

Hierarchical segmentation-based software for cover classification analyses of seabed images (Seascape)

“…Ewing et al 1967, Torlegård & Lundalv 1974, Corliss et al 1979, Gutt et al 1996, Clarke 2003, Bailey et al 2007, Lirman et al 2007, Bowen et al 2009, Shortis et al 2009). This includes the use of manual image equipment (digital still and video cameras) operated by SCUBA divers and remote imaging technology in deep waters, such as remotely operated vehicles (ROVs), autonomous underwater vehicles (AUVs), landers equipped with cameras for long-term observations and a variety of other still camera systems.…”

Efficient Topology Estimation for Large Scale Optical Mapping

“…4.7). This second dataset was composed of 30 images, extracted from an underwater image sequence acquired by a Phantom 500 ROV during a survey in Andros, in the Bahamas [66]. The results for the second dataset are summarised in Table 4.3.…”

“…1.2 and 1.3). Many scientifically interesting sites are located in areas which are nearly flat, such as the coral reefs in the Florida Reef Tract [66]. We consider in this thesis cases where the 3D relief of the scene is negligible compared to the altitude of the robot, and the seafloor is therefore assumed to be and is modelled as a planar scene 1 .…”

“…Therefore, mosaicing techniques are needed to create high-resolution maps of the surveyed area using a large number of acquired images and to get a global perspective of the underwater terrain [45,89,109,60,94,96]. Thus, robotic exploration with the aim of constructing photo-mosaics is becoming a common requirement in geological [112,33] and archaeological surveys [36], mapping [59], ecology studies [56,66,89], environmental damage assessment [41,65] and temporal change detection [29]. Owing to the rapid development in data acquisition platforms, there is an increasing need for large-scale image mosaicing methods.…”

Reconfigurable AUV for intervention missions: a case study on underwater object recovery

“…The mosaic building process is detailed in [16,17]. It starts by searching for correspondences between consecutive images (referred to as consecutive image registration) to determine their homographies.…”