The platform is designed to support scientists and investigators in academic institutes, environmental agencies, public bodies as well as environmental consultancy companies. The services for management and analysis of marine data are open-source and can be used by anyone. If you want to upload and analyse your footage on SUBSIM, please contact us through the SBDI Support Centerand we will help you to get started.
SUBSIM’s modular architecture
SUBSIM currently consists of five components, each with a distinct function explained in the below table. The components indicated with an asterisk (*) are making use of the Swedish National Infrastructure for Computing (SNIC), which are only accessible to academic users. Non-academic users will be directed to alternative server space.
SUBSIM high-level workflow
When you use SUBSIM, you typically start with organising your media and metadata and the upload these to the Cloudina server, see (a) in the below figure. Thereafter you can label the uploaded images and movies, with or without help from citizens and peers using the Zooniverse platform (b). Once you have created training data, you can train machine-learning models to perform e.g., object detection, image classification, or image segmentation (c). Please note, the image segmentation function is not yet fully implemented. Once your model is well trained and evaluated, it can be published and deployed to extract object and image classes from your media on a high-perpromance computing cluster (d). Finally, there is also an option to publish the machine-based species observations to the Global Biodiversity Information Facility, GBIF (e).
SUBSIM GitHub repo
This is the main development branch for SUBSIM. Here you find all source code and collaborators of the SUBSIM project.
SEafloor Annotation and Mapping Support (SEAMS) App
The SEAMS App, is a tool designed for the interpretation of seafloor images and video. Developed for the Swedens Geological Survey (SGU) this application streamlines the annotation and mapping process of underwater data using the visuel methods.
Create raster maps from model detections
This repo contains scripts and example files to convert SUBSIM model detections into GIS layers. It features custom-made R functions built around the SUBSIM YOLO object detection models output and Python scripts for geoprocessing in QGIS (PyQGIS).
SBDI Github repo
Here you find the repo for the development of the Swedish Biodiversity Data Infrastructure (SBDI), to which SUBSIM is connected and delivers data to.