内容来源：https://www.sciencenews.org/article/ai-tech-save-gray-whales

内容总结：

新型AI热成像系统护航旧金山湾，拯救灰鲸免受船只撞击

在美国加利福尼亚州繁忙的旧金山湾，每年都有大量灰鲸因误入航道而丧命于船只撞击。为应对这一日益严峻的威胁，一个由海洋科学家、美国海岸警卫队、鲸鱼追踪专家及当地渡轮公司组成的联合团队，于5月19日正式启动了一套基于人工智能的新型监测系统。

这套系统的核心在于将部署在海湾不同位置的24小时热成像摄像头与AI技术相结合。当AI识别出可能远至7公里外的鲸鱼热信号（鲸鱼喷出的水柱或身体温度比周围海水高出约2摄氏度）后，专家会立即进行人工核实，确认后随即向周边船只发出警报，敦促其减速或改变航线，以避免致命碰撞。

这项技术由马萨诸塞州伍兹霍尔海洋研究所的科学家研发，并成立了名为WhaleSpotter的公司进行推广。项目首席科学家丹尼尔·齐特巴特表示，团队耗费约15年时间，使用数十万张热成像训练AI，以实现远距离精准识别，为大型集装箱船这类惯性大、难以及时转向的船舶争取宝贵的反应时间。

根据规划，该系统已在天使岛的一座无线电塔上安装摄像头，用于监控多条繁忙的航运路线。第二台摄像头将安装在一艘每日往返的渡轮上，未来还计划在金门大桥和恶魔岛等地增设监测点。

这一举措的背景是灰鲸正面临严重的生存危机。自2018年以来，研究者观察到灰鲸在从阿拉斯加向南墨西哥的1.6万公里长途迁徙中，越来越多地“绕道”进入旧金山湾觅食。科研人员分析，这是由于气候变化导致北极海冰融化，破坏了灰鲸主要食物来源（生活在海底沉积物中的端足类动物）的生存环境，迫使饥饿的鲸鱼冒险进入人类活动密集的海湾寻找食物。

数据显示，灰鲸种群数量已从2018年的约2.05万头锐减至2023年的1.45万头。仅2025年，旧金山湾及其周边就发现21头灰鲸尸体，其中五分之二的死因是船只撞击。加州大学圣塔芭芭拉分校海洋生态学家道格拉斯·麦考利指出：“看着这些饥饿的鲸鱼在繁忙的旧金山湾中跌跌撞撞，令人心碎。每一天都令人揪心……这套新系统将拯救鲸鱼的生命。”

当地鲸类生物学家约瑟芬·斯拉托表示，她对此“谨慎乐观”，并欣慰地看到航运业等多个部门和机构正合作开发基于科学的长期解决方案。齐特巴特强调：“航运不会消失，我们需要一项技术，既能让我们利用海洋，也能让鲸鱼安稳生活。我们希望实现尽可能多的部署，这意味着我们在海洋上拥有更多双‘眼睛’。”

中文翻译：

一款由人工智能驱动的监测系统，或将拯救那些因致命绕行而日益频繁闯入加州交通繁忙的旧金山湾的灰鲸生命。
这项新技术将部署在湾区不同地点的全天候热成像摄像头与人工智能相结合，能够探测到远在7公里外的鲸鱼。一旦科学家确认探测到鲸鱼，系统便会发出警报，警告该区域的船只减速或改变航向以避免碰撞。
由海洋科学家、美国海岸警卫队、鲸鱼追踪专家及当地渡轮公司组成的联盟于5月19日宣布在湾区部署该系统。一台安装在湾区天使岛无线电塔上的摄像头将监控多条繁忙的航运路线。第二台摄像头将安装在一艘每日穿行海湾的客运渡轮上，未来可能的额外摄像头安装点包括金门大桥和恶魔岛。
这项基于人工智能的鲸鱼探测技术源于马萨诸塞州伍兹霍尔海洋研究所（WHOI）研究人员的创意，他们随后成立了名为WhaleSpotter的公司来推广该技术。WHOI物理学家、WhaleSpotter首席科学家丹尼尔·齐特巴特表示：“我们希望能在极远处探测到鲸鱼，从而让船员有时间采取行动。”这对于大型船舶，例如惯性巨大、无法迅速改变航向的集装箱船来说尤为重要。
齐特巴特称，开发可靠的鲸鱼探测系统耗时约15年。鲸鱼喷出的水柱或鲸鱼本身的体温比周围海水高出约2摄氏度。因此，研究人员使用了数十万张热成像图像来训练人工智能识别这些相对温差，从而判定鲸鱼的存在。随后，一旦系统发出探测信号，WhaleSpotter的研究人员将核实数据，以最大限度减少误报。经核实后，警报会发送给附近所有船只。
齐特巴特表示：“我们希望部署尽可能多的系统，因为这最终意味着我们能更全面地观察海洋。航运不会消失。我们需要一种技术，既能让我们利用海洋，又能让鲸鱼正常生活。”
2025年，在旧金山湾及其周边地区发现了21头灰鲸的尸体；研究人员称，其中五分之二的死亡是由船只撞击造成的。这些死亡案例是研究人员自2018年首次观察到的一个令人担忧的趋势的一部分：鲸鱼在从阿拉斯加海岸的觅食地向南迁徙至墨西哥附近的交配地，全程长达1.6万公里的路途中，越来越多地选择在湾区停留。
这些鲸鱼很可能处于饥饿状态。在北极，它们以沉积在海洋沉积物中的名为端足类的小型甲壳动物为食；而这些端足动物则依靠生长在海冰下层的藻类为生。气候变化正在迅速融化海冰，从而破坏了食物链。
灰鲸种群数量从2018年的大约20500头急剧下降到2023年的约14500头。数百头鲸鱼被发现搁浅在北美西海岸，其中许多鲸鱼营养不良。因此，为了在剩余的迁徙途中维持生存，它们便进入湾区寻找食物。
加州大学圣塔芭芭拉分校的海洋生态学家道格拉斯·麦考利在5月19日的一份新闻稿中说：“看到这些饥饿的鲸鱼在喧嚣忙碌的旧金山湾中茫然徘徊，令人心碎。”麦考利是该校贝尼奥夫海洋科学实验室的主任，该实验室是开发并部署这项新技术的联盟合作伙伴之一。“每一天都令人提心吊胆……这套新系统将拯救鲸鱼的生命。”
加州罗内特公园的索诺玛州立大学鲸鱼生物学家约瑟芬·斯拉托表示，她希望这项技术能成为“保护旧金山湾鲸鱼的正确方向上的巨大飞跃”。
斯拉托说：“我持谨慎乐观的态度。我非常高兴船只撞击问题正得到认真对待。”她补充说，尤其令人鼓舞的是，包括航运业在内的如此多不同的组织和合作伙伴正在共同努力，寻求基于科学的长期解决方案。

英文来源：

An AI-powered monitoring system could save the lives of gray whales that are increasingly taking a deadly detour into California’s heavily trafficked San Francisco Bay.
The new technology combines round-the-clock thermal cameras deployed at different locations in the bay with AI to detect whales that may be as far as 7 kilometers away. Once the whale detection is confirmed by scientists, an alert goes out to warn vessels in the area to slow down or change course to avoid a collision.
A coalition of ocean scientists, the U.S. Coast Guard, whale tracking experts and local ferry companies unveiled the deployment in the bay on May 19. A camera mounted on a radio tower on Angel Island within the bay will monitor numerous busy shipping routes. A second camera will be installed on a passenger ferry that crosses the bay daily, and future additional camera sites could include the Golden Gate Bridge and Alcatraz.
The whale-detecting AI-powered tech is the brainchild of researchers at the Woods Hole Oceanographic Institution, or WHOI, in Massachusetts, who later created a company called WhaleSpotter to market the tech. “We wanted to be able to detect whales so far out that it would give mariners time to take action,” says Daniel Zitterbart, a physicist at WHOI and the chief scientist of WhaleSpotter. That’s particularly important for large ships, such as container vessels, that have a great deal of inertia and can’t quickly change course.
Developing a reliable whale detection system took about 15 years, Zitterbart says. Water emitted from whales’ blowholes, or the whales’ bodies themselves, is warmer than the ambient water by about 2 degrees Celsius. So the researchers used hundreds of thousands of thermal images to train the AI to recognize those relative temperature differences as signifying a whale. Then, when there’s a detection, a WhaleSpotter researcher will verify the data, to minimize false positives. Once verified, an alert is sent to any vessels nearby.
“We want as many deployments as possible, because that ultimately means we have better eyes on the ocean,” Zitterbart says. “Shipping is not going to disappear. We need to have a tech that allows us to use the ocean, but also allows the whales to go about their lives.”
In 2025, 21 gray whales (Eschrichtius robustus) were found dead in and around San Francisco Bay; two-fifths of those deaths were due to ship strikes, researchers say. The deaths are part of a disturbing trend that researchers first observed in 2018: The whales were increasingly making a pit stop in the bay along their 16,000-kilometer-long migration southward from their feeding grounds off Alaska’s coast to their mating grounds near Mexico.
The whales were likely hungry. In the Arctic, they feed on tiny crustaceans called amphipods in ocean sediments; those amphipods, in turn, are nourished by algae that grows on the underside of sea ice. Climate change is rapidly melting that sea ice, disrupting the food chain.
Gray whale populations declined dramatically from about 20,500 in 2018 to about 14,500 in 2023. Hundreds of whales were found stranded along the North American west coast. Many of those whales were suffering from malnutrition. So, to sustain themselves for the rest of their migration, they have been heading into the bay looking for food.
“It is heartbreaking to see these starving whales stumbling around in the middle of the hustle and bustle of San Francisco Bay,” University of California, Santa Barbara marine ecologist Douglas McCauley said May 19 in a news release. McCauley is the director of UCSB’s Benioff Ocean Science Laboratory, one of the coalition partners that developed and is deploying the new technology. “Every day is a nail-biter.… This new system will save whales’ lives.”
Josephine Slaathaug, a whale biologist at Sonoma State University in Rohnert Park, Calif., says she hopes this technology will be a “huge leap in the right direction to protecting whales in San Francisco Bay.”
“I’m cautiously optimistic,” Slaathaug says. “I’m very glad that the vessel strike issue is being taken seriously.” And it’s especially heartening, she adds, to see so many different organizations and partners — including the shipping industry — working together to develop a science-based, long-term solution.