Automated Image Analysis for Rapid Biostratigraphic Data Collection

钙质超微化石生物地层学是勘探和生产过程中的一个关键工具，提供了跨领域和跨领域的强有力的高分辨率地层对比。在钻井作业中，实时的钻井现场生物地层学对于地质导向和地质停止活动都是必不可少的。这些工作目前依赖于用标准光学显微镜技术密集分析纳米化石的内容。本专业将在显微镜和图像捕捉自动化的最新发展的基础上，结合“智能”图像处理和分类算法，开发一种新的自动纳米化石组合数据收集系统。该项目将专注于使用能够在颗粒分类、识别和形态分析方面具有高技能的图像处理算法来实现高通量的自动图像捕获。首先，该项目将专注于连续的碎屑和碳酸盐沉积序列，跨越过去1000万年，从澳大利亚西北部大陆架上的Browse盆地恢复。这些沉积物产生了很好的钙质超微化石恢复和保存，并且已经有了很好的古地磁和浮游有孔虫年龄控制，可以作为拟议的新生物地层学研究的基础。该项目直接建立在首席主管Dunkley Jones的研究基础上，Dunkley Jones在纳米化石生物地层学方面拥有约15年的经验。该项目将由斯蒂芬·劳滕施莱格博士和Ales Leonardis教授支持，前者专注于脊椎动物的3D数字成像和修复，后者是计算机科学学院的计算机视觉专家。我们还与项目合作伙伴Manuel Vieira博士密切合作，他将就与行业相关的应用提供直接指导，并将其与壳牌在远程自动化显微镜平台现场应用方面的现有实验联系起来。

Calcareous nannofossil biostratigraphy is a key tool within the exploration and production process, providing robust high-resolution stratigraphic correlations across and between fields. Within drilling operations, real-time rig-site biostratigraphy can be essential for both geo-steering and geo-stopping activities. These operations currently rely on intensively analysing nannofossil content with standard light microscopy techniques. This studentship will build on recent developments in microscope and image capture automation, together with "smart" image processing and classification algorithms, to develop a new system for automated nannofossil assemblage data collection. This project will focus on high- throughput automated image capture using image-processing algorithms capable of high-skill in particle classification, identification and morphometric analysis. In the first instance the project will focus on a continuous sequence of mixed clastic and carbonate sediments, spanning the last 10 million years, recovered from the Browse Basin on the NW Australian shelf. These sediments yield excellent calcareous nannofossil recovery and preservation and already have good paleomagnetic and planktonic foraminiferal age control within which to situate the proposed new biostratigraphic study. This project directly builds on the research of lead supervisor Dunkley Jones, who has ~15 years of experience in nannofossil biostratigraphy. The project will be supported by Dr Stephan Lautenschlager whose research focuses on the 3D digital imaging and restoration of vertebrates, and Prof. Ales Leonardis, an expert in computer vision with the School of Computer Sciences. We also work closely with project partner, Dr Manuel Vieira, who will provide direct guidance on industry-relevant applications, and make the link to Shell's existing experiments in remote, automated microscopy for rig-site applications.