LeisAIR: Leisure watercraft AI Image Recognition

NASH Maritime是航运，导航和海上风险的专家。为了管理安全和最大限度地提高港口和海港的容量，必须通过航行风险评估（NRA）流程识别、测量和减轻风险。NRA是海事安全管理的关键部分，也是任何海事基础设施项目同意过程中的重要步骤。NRA不可或缺的是船舶交通数据分析。通过自动识别系统（AIS），可以随时获得商业和一些大型娱乐船只的鲁棒，可靠的位置和运动数据。然而，没有关于娱乐用水者-小船、橡皮艇、赛艇运动员、皮划艇运动员等-的此类数据。相反，关于娱乐用水者的运动和水量的数据是通过利益攸关方协商和直接观察收集的。前者非常主观，后者需要大量资源，而且时间有限。拥有娱乐用水用户的准确数据将使NRA过程的准确性提高一个数量级，促进使用适当的风险控制，使多个海事利益相关者和谐共存，并最大限度地提高所有水用户的能力。NASH海事和布鲁内尔大学将利用人工智能图像识别能力的最新进展来自动化娱乐船舶数据收集过程。该项目将基于NASH Maritime现有的人工智能模型，以提高和扩展其准确识别和记录港口和港口闭路电视信号中各种不同类型的娱乐船只数量的能力。该模型将用于提高准确性并降低NRA的交付成本。这将在交付导航风险评估的准确性方面实现一个台阶式的变化，并使该过程更具成本效益，最大限度地提高海上运输能力并提高海上安全。该技术的应用范围超出了NRA，NASH将探索使用娱乐船只图像识别来支持：* 其他海上作业的风险评估 * 用水管理/绘图研究 * 执行当地法规 * 安全 * 航行决策支持

NASH Maritime are specialists in shipping, navigation and maritime risk. To manage safety and maximise capacity in ports and harbours, it is essential that risks are identified, measured and mitigated through a process of Navigation Risk Assessment (NRA). The NRA is a key part of the management of maritime safety and an essential step in the consenting process of any maritime infrastructure project.Integral to NRA is the analysis of vessel traffic data. Robust, reliable location and movement data for commercial, and some large recreational, vessels are readily available via the Automatic Identification System (AIS). However no such data are available for recreational water users -- small craft, dinghies, rowers, kayakers etc.Instead, data on the movements and volumes of recreational water users is gathered through stakeholder consultation and by direct observation. The former is very subjective, the latter resource intensive and time-limited. Having accurate data on recreational water users would improve the accuracy of the NRA process by an order of magnitude, promoting the use of appropriate risk controls, enabling the harmonious coexistence of multiple maritime stakeholders and maximising capacity for all water users.NASH Maritime and Brunel University will use recent advances in the capability of AI image recognition to automate the process of recreational vessel data collection The project will build on NASH Maritime's existing AI model to improve and extend its capability to accurately identify and record volumes of various different types of recreational craft from port and harbour CCTV feeds. The model will be used to increase the accuracy and reduce the cost of delivering NRAs.This will deliver a step-change in the accuracy of delivering navigation risk assessments and make the process more cost-effective, maximising maritime transport capacity and improving maritime safety.There are applications for this technology beyond NRA, and NASH will explore the use of recreational vessel image recognition to support:* Risk assessment for other maritime operations* Water use management/ mapping studies* Enforcement of local regulations* Security* Navigation decision support