Towards Autonomous 3D Scanning of Underwater Environments
PhD Student, CSSE
Time & Place
Thu, 06 May 2021 12:00:15 NZST in JE443
Low cost imaging sensors are enabling mass data collection of environments and habitats for monitoring, as well as real-time species identification for farming.
The New Zealand cryptic scallop Pecten Novaezelandiae is a valuable food resource, however, current dredge harvesting practices have damaged benthic habitats to the point of multiple fisheries closures which could span almost a decade, negatively impacting recreational, customary, and commercial fishing. Applying deep learning, specifically convolutional neural networks to in situ identification and sizing will allow for better monitoring of fisheries, as well as the development of better harvesting practices- such as selective minimal impact robotic harvesting. Heterogenous stereo vision, robust visual keypoints, autonomous 3D scanning, and efficient CNN implementations are the subjects of this PhD.
I'm Tim Rensen, a PhD student funded by NIWA supervised by Dr Oliver Batchelor and Prof. Richard Green. Undergraduate degrees in Mechatronics Engineering and Physics. I love getting outdoors and am driven to progress perception and understanding of the natural world around us to improve the symbiotic relationships with our world.