AI-based seagrass morphology measurement.

Seagrass meadows are an essential part of the Great Barrier Reef ecosystem, providing various benefits such as filtering nutrients and sediment, serving as a nursery for fish and shellfish, and capturing atmospheric carbon as blue carbon. Understanding the phenotypic plasticity of seagrasses and their ability to acclimate their morphology in response to environ-mental stressors is crucial. Investigating these morphological changes can provide valuable insights into ecosystem health and inform conservation strategies aimed at mitigating seagrass decline. Measuring seagrass growth by measuring morphological parameters such as the length and width of leaves, rhizomes, and roots is essential. The manual process of measuring morphological parameters of seagrass can be time-consuming, inaccurate and costly, so researchers are exploring machine-learning techniques to automate the process. To automate this process, researchers have developed a machine learning model that utilizes image processing and artificial intelligence to measure morphological parameters from digital imagery. The study uses a deep learning model called YOLO-v6 to classify three distinct seagrass object types and determine their dimensions. The results suggest that the proposed model is highly effective, with an average recall of 97.5%, an average precision of 83.7%, and an average f1 score of 90.1%. The model code has been made publicly available on GitHub (https://github.com/sajalhalder/AI-ASMM).

Vulnerability and Impact Analysis of the IEC 61850 GOOSE Protocol in the Smart Grid.

IEC 61850 is one of the most prominent communication standards adopted by the smart grid community due to its high scalability, multi-vendor interoperability, and support for several input/output devices. Generic Object-Oriented Substation Events (GOOSE), which is a widely used communication protocol defined in IEC 61850, provides reliable and fast transmission of events for the electrical substation system. This paper investigates the security vulnerabilities of this protocol and analyzes the potential impact on the smart grid by rigorously analyzing the security of the GOOSE protocol using an automated process and identifying vulnerabilities in the context of smart grid communication. The vulnerabilities are tested using a real-time simulation and industry standard hardware-in-the-loop emulation. An in-depth experimental analysis is performed to demonstrate and verify the security weakness of the GOOSE publish-subscribe protocol towards the substation protection within the smart grid setup. It is observed that an adversary who might have familiarity with the substation network architecture can create falsified attack scenarios that can affect the physical operation of the power system. Extensive experiments using the real-time testbed validate the theoretical analysis, and the obtained experimental results prove that the GOOSE-based IEC 61850 compliant substation system is vulnerable to attacks from malicious intruders.