Grant: Engage Grants Program

The objective of this project is to develop an image mosaic system to integrate with the underwater camerasdeveloped by SubC Control Ltd. in Clarenville, Newfoundland. SubC is developing video technology forsubsea survey applications using autonomous or remotely operated underwater vehicles. This technologyallows a user to gather high definition video of surveys such as sea bed mapping, spice identification, marinelife research, wreckage assessment etc. The current SubC product suit is only capable of providing the rawvideo feed, leaving the client to seek support from a third party for post processing of this data. Therefore,SubC is interested to perform additional R&D to enhance its product line, mainly in the area of enablingembedded computer vision features. These computer vision capabilities include image mosaicking, point cloudgeneration, mapping, object identification, tracking and classifications. As part of this initiative, this researchwork aims to design a fully functional image mosaic system to integrate with the current imaging systemdeveloped by SubC to enhance its real time raw data processing capability and provide user with aggregatedinformation for storage or planning operations. Key challenges in image mosaic process include developmentof a global image topology for entire area of interest, determination of a fast and computationally efficientalgorithm for image registration and optimization to generate an optimal global image mosaic. In the proposedmethod the AUV sensor data will be integrated with the video data for efficient image topology generation andthe historical data of underwater surveys will be used to optimize the feature based image registration process.As part of validation, the developed system will be integrated with the SubC imaging hardware and will betested in the SubC lab facility in NL.