OFI SF5 Advanced capacity for autonomous detection of green sea urchin (Strongylocentrotus droebachiensis)

在过去的5年里，Gagnon博士和他的行业合作伙伴已经证明，纽芬兰的绿色海胆(Strongyloctrotus Droebachiens)性腺产业既是可以实现的，也是非常可取的。迫切需要开发新的技术来快速有效地收集绿色海胆，需要更多的信息来可持续地管理海胆种群及其栖息地。遥控潜水器(ROV)提供了这一潜力，但目前还没有用于这一特定目的的商用ROV解决方案。为了利用Gagnons博士之前的研究创造的势头，拟议的OFI种子基金项目旨在发展先进的ROV能力的基础，以便在自然环境条件和栖息地特征下智能、半自主地识别和收集海胆。这个多学科项目将建立基于图像识别和机器学习的海胆识别能力。将通过开发一套集成的计算机程序来解决这一问题，该程序可以识别使用标准和3-D潜水高清和激光摄像系统获取的海底视频中的海胆并对其进行电子标记，并通过自学能力逐步提高海胆识别能力。该项目将分配给一位具有人工神经网络和计算机视觉相关背景的顶尖博士生。它将作为一个跳板，使纪念大学不同学科的三名研究人员及其来自NRC和Kraken Robotics(行业合作伙伴)的同事能够就这一新倡议进行合作，这一新倡议为专利/商业化和最终纳入其他海洋物种提供了强大的潜力。

Over the last 5 years, Dr. Gagnon and his industry partners have shown that a Newfoundland green sea urchin (Strongylocentrotus droebachiensis) gonad industry is both achievable and highly desirable. There is a pressing need to develop new technologies to quickly and efficiently collect green sea urchins, and more information is needed to sustainably manage sea urchin stocks and their habitats. Remotely operated vehicles (ROV) offer this potential, yet there is currently no commercially available ROV solution for this specific purpose. To capitalize on momentum created by Dr. Gagnons previous research, the proposed OFI Seed Fund project aims to develop the foundation of advanced ROV capacity for intelligent, semi-autonomous recognition and collection of urchins under natural environmental conditions and habitat characteristics. This multidisciplinary project will build sea urchin identification capacity based on image recognition and machine learning. It will be addressed by developing an integrated suite of computer programs that can recognize and electronically tag urchins on videos of seabed acquired with standard and 3-D submersible HD and laser camera systems, and progressively improve at urchin recognition via self-learning capacity. The project will be assigned to a top-notch PhD candidate with relevant background in Artificial Neural Networks and Computer Vision. It will act as a springboard enabling three researchers from different disciplines at Memorial University and their colleagues from NRC and Kraken Robotics (industry partner) to collaborate on this new initiative that presents a strong potential for patenting/commercialization and eventual inclusion of other marine species.