Innovative AI Solution Shows Initial Success in Protecting Chinese White Dolphins

Innovative AI Solution Shows Early Success in Protecting Chinese White Dolphins

In a groundbreaking move, Huawei and its partners announced the preliminary results of an innovative artificial intelligence (AI) solution aimed at studying and protecting the endangered Chinese white dolphins in Xiamen Bay, China. This initiative, launched three months ago as part of Huawei's TECH4ALL program, has already yielded promising results, making considerable headway in the conservation efforts for this iconic species.

Over the past few months, the AI system has successfully identified 13 individual dolphins from an extensive dataset of 2,820 images and videos. This identification was achieved with an impressive accuracy rate of over 90% and complex behavior recognition reaching 85%. Furthermore, the system improved data labeling efficiency by an astounding 400% and enhanced the response time of law enforcement officials by 65% towards vessels breaching speed limits or invading protected areas.

Cui Yangyang, the director of Huawei’s TECH4ALL Program Office, emphasized that the data analytics generated by the AI is pivotal for conservationists in formulating relevant protection measures tailored to understand and tackle the threats facing this vulnerable species, ensuring their long-term survival and prosperity in the wild. The Chinese white dolphin is currently regarded as a vulnerable species, with significant populations inhabiting Xiamen Bay, which accounts for 51% of China’s dolphin population.

The challenges these dolphins face are numerous, including increased maritime traffic, fishing activities, coastal engineering projects, habitat loss, and pollution. Sounds from ships can endanger dolphin populations, leading to entanglements in fishing gear and diminishing their natural habitats, all of which threaten their survival.

To combat these threats, Huawei is collaborating closely with the Third Institute of Oceanography under the Ministry of Natural Resources of China, as well as China Mobile. The project includes a unique feature recognition system capable of identifying each dolphin based on the unique markings on their dorsal fins. Wang Xianyan, head of the Marine Endangered Species Research and Conservation Team at the Third Institute of Oceanography, pointed out that the incorporation of AI has made studies more efficient and conservation efforts more focused. The AI recognition system provides crucial data on individual survival rates, reproductive dynamics, and social interactions, which are essential for formulating effective conservation strategies.

The AI solution utilizes a combination of image preprocessing, AI-enabled recognition, dorsal fin image cropping, data classification, and visualization through cloud technology. Each dolphin has its own file allowing researchers to monitor the health and condition of each individual dolphin. Such long-term data is vital for developing protection measures and monitoring population dynamics, including the number, distribution, age structure, reproductive behaviors, and threats faced by the dolphin population.

Previously, manual monitoring methods were labor-intensive and struggled to provide the reliable information necessary for tracking population dynamics or creating effective conservation measures. The Third Institute of Oceanography estimates that the next 10 to 15 years will be crucial for the growth of the Chinese white dolphin population, which plays an integral role in coastal marine ecosystems and contributes to carbon cycling through its feeding habits, impacting plankton dynamics.

In tandem with AI analytics, a 5G-A network comprising 10 base stations covers an expansive area of 330 square kilometers of Xiamen Bay, ensuring comprehensive coverage of key conservation zones. This advanced communication system integrates detection capabilities with real-time monitoring through radar, visual terminals, satellites, and vessel automatic identification systems (AIS). Such innovations enable tracking of vessel movements within a 20-kilometer radius, allowing for prompt AI-generated alerts for breaches like speed violations or restricted area incursions. Law enforcement can monitor these infractions from land, facilitating rapid response to offenders. So far, 12 vessels have been investigated for posing potential risks to the dolphin population.

The ongoing research utilizing AI is gradually unveiling essential aspects of Chinese white dolphins' lives in Xiamen Bay, offering hope for the future of this threatened species. The intricate interplay of technology and conservation efforts represents a beacon of hope for both the dolphins and marine ecosystems as a whole.