Figure 1: Schematic overview of the proposed image data workflow from acquisition through curation and management. | Scientific Data

Figure 1: Schematic overview of the proposed image data workflow from acquisition through curation and management.

From: An acquisition, curation and management workflow for sustainable, terabyte-scale marine image analysis

Figure 1

Various robots (autonomous underwater vehicles (AUVs), landers, remotely operated vehicles (ROVs), towed platforms) create stacks of imagery (a) and metadata tables (b). Erroneous metadata values (here marked in red) and corrupt imagery (e.g. black images where the flash did not fire) might occur. Metadata are attached to the image data, image processing is applied and corrupt and erroneous data are flagged and filtered out (c). The resulting curated data set is the quality controlled data product that is suitable for publication and analysis. Metadata and image data are stored in suitable databases (public or private). Image data items should be linked to their corresponding metadata at archiving. The individual steps from pre-cruise planning to publication are discussed in the text. For a specific use case, see Fig. 3.

Back to article page