近日，研究人员 David Kroodsma 联合美国国家地理（National Geographic）及彭博社慈善基金会（Bloomberg Philanthropies）利用卫星测绘数据及 AI 技术建立了世界首款“全球船舶足迹地图”。这款地图号称能够准确记录人类在海洋中的各种动向，弥补了传统船舶识别系统的不足。
据悉，目前业界主要通过 VMS（船舶监视系统）及 AIS（船舶自动识别系统）监测全球船舶动向，但相关系统存在“设备装载基数”及“天气情况”等桎梏，难以准确反映全球船舶动向。而 David Kroodsma 主要利用 AI 识图形式，训练了 3 款用于检测卫星图像的 AI 模型，并利用 2017 年 -2021 年间的全球卫星数据建立了相对全面的“海上船舶足迹地图”。
研究表明，全球平均有 6.33 万艘船只在同时航行，但其中近 75% 的渔船及近 25% 的各式船只实际上没有出现在 VMS 及 AIS 船舶检测系统中，这表明传统船舶识别系统无法准确追踪全球船舶动向情况。此外，研究人员还利用 AI 对全球海上能源基础设施进行了盘点，其中全球最大的海上石油基础设施集中在北美墨西哥湾，美国拥有其中近 2200 座，占全球海上石油基础设施的 1/4。在海上风电方面，2017 年至 2022 年，我国风力涡轮发电机数量增长了 900%，增速世界第一，平均每年新建 950 台。
通过这款全球船舶足迹地图，研究人员为国际社会提供了更为精准的海洋动态信息，有助于推动海洋资源保护和海洋环境监测。
英文翻译：
News Title: Global Ship Tracks Map Unveiled with AI Technology
Keywords: Global Ship Tracks, AI Technology, Ocean Monitoring
News Content:
Recently, researcher David Kroodsma has collaborated with National Geographic and Bloomberg Philanthropies to develop the world’s first “Global Ship Tracks Map” using satellite mapping data and AI technology. This map claims to accurately record various human activities in the ocean, filling the gap in traditional ship identification systems.
It is known that currently, the industry mainly monitors global ship movements through VMS (Vessel Monitoring System) and AIS (Automatic Identification System). However, these related systems have limitations such as “equipment loading base” and “weather conditions,” making it difficult to reflect the global ship movement accurately. David Kroodsma mainly uses AI image recognition to train three AI models for detecting satellite images and builds a relatively comprehensive “ocean ship tracks map” with global satellite data from 2017 to 2021.
The study shows that there are an average of 63,000 ships sailing simultaneously worldwide, but nearly 75% of fishing boats and nearly 25% of various ships are not actually detected by VMS and AIS ship monitoring systems. This indicates that traditional ship identification systems cannot accurately track global ship movements. In addition, researchers have also used AI to inventory global offshore energy infrastructure. The largest offshore oil infrastructure is concentrated in the North American Gulf of Mexico, where the United States has nearly 2,200 installations, accounting for 1/4 of global offshore oil infrastructure. In terms of offshore wind power, the number of wind turbines in China increased by 900% from 2017 to 2022, ranking first in the world with an average of 950 new installations each year.
Through the Global Ship Tracks Map, researchers provide more accurate ocean dynamic information to the international community, contributing to the protection of ocean resources and marine environment monitoring.
【来源】https://www.ithome.com/0/743/450.htm
Views: 1