Seaseer-project Description for SoC 2026

pages/development/google-summer-of-code/2026.md

@@ -600,6 +600,34 @@ The platform is intentionally scoped as a learning-oriented reference implementa
 
 <hr>
 
+### SeaSee'r: Underwater Mapping and Exploration using Spatially anchored Panoramas
+
+350 Hours  
+{: .label .label-blue }
+
+{: .highlight }
+> **Required Skills**: Web Development (Full Stack), Web 3D Graphics, Real-Time Rendering, Python, Computer Vision, Git, REST APIs  
+> **Possible Mentors**: Benedikt Kantz, Tobias Schreck  
+> **Expected Outcome**: A navigatable underwater map based on real multimodal data from a remotely operated vehicle (ROV), including panorama (videos) and sonar data
+> **Difficulty level**: Advanced  
+> **Project size**: Large  
+> **Task link**: [View task](https://docs.google.com/document/d/1jVWFRFLTccSHU0lVX65LY1HiH_VQi3quhmlc7PmYuC8/edit?usp=sharing)
+
+Project Description
+
+Modern underwater ROVs provide a magnitude of different data streams in both real-time and post-operation for scientific application. There is, however, a lack of systems integrating the domain-specific requirements for navigating and exploring these data, especially for frequent site visits to study the behavior of marine creatures.
+
+The project therefore proposes the creation of an open-source system to ingest data from the ROV and display the panorama images spatially anchored based on the position and orientation of the vehicle at the specific time stamps. The resulting interface should also incorporate multimodal data, i.e. sonar data and possible further data streams like point clouds directly into the visual exploration system. Once this basic visualization system is set up, specific exploration tasks can be implemented, such as:
+
+- Mission and path planning, with
+- Search for prior exploration passes for detailed mission planning at specific sites,
+- Point cloud stitching / Image registration to compare passes over the same regions,
+- Site Re-Identification even in noisy environments through multimodal data use, and possible
+- Registration of new images and feeds live (during missions) to improve the navigation accuracy.
+
+The project serves as an exploratory foray into bringing advances in spatial exploration and mapping techniques below the sea, and requires an extensible and longevity-focused architecture to enable downstream additions of data modalities and different vehicle types. The specific mission planning aspects need to be flexible in their structure as well, to allow for a rapid evolution in supported planning tooling; as the scientific goals might shift over time -- or new aspects of the marine life are discovered, requiring new paradigms.
+
+<hr>
 
 ### Your own Project Ideas ...
 90, 175 or 350 Hours