Skip to main content

Huawei's AI solution enhances Chinese white dolphin protection

Huawei and partners announced preliminary results of an AI-powered solution aimed at studying and protecting the Chinese white dolphins in Xiamen Bay. This initiative, part of the Huawei TECH4ALL program, commenced three months ago and achieved notable advancements in dolphin identification and conservation monitoring.

The project identified 13 individual dolphins from a dataset of 2,820 images and videos, attaining an individual identification accuracy exceeding 90%. Furthermore, it reported a recognition rate of complex behaviors around 85% and improved the efficiency of data labeling by 400%. The response time of law enforcement to ships entering conservation areas also increased by 65%.

Cui Yangyang, Director of the TECH4ALL Program Office for Huawei, stated that the insights garnered through Artificial Intelligence (AI) facilitate the development of targeted conservation measures to protect this species. The Chinese white dolphin is classified as vulnerable under the IUCN Red List and faces challenges in Xiamen Bay, where over half of China’s dolphin population resides.

The partnership with the Third Institute of Oceanography and China Mobile included a comprehensive feature recognition system for identifying dolphins based on distinctive dorsal fin markings. Wang Xianyan from the Marine Endangered Species Research and Conservation Team noted that the AI system generates essential data on survival, reproductive dynamics, and social interactions, aiding in effective conservation strategies.

The solution encompasses image preprocessing, AI inference, dorsal fin cropping, data classification, and cloud-based display, enabling researchers to track each dolphin's status. With long-term data crucial for protection measures, this approach enhances understanding of population dynamics and threats faced by the species.

According to projections from the Third Institute of Oceanography, the next 10 to 15 years are critical for the growth of the white dolphin population, which plays a key role in marine ecosystems and carbon cycling. Protecting these dolphins contributes to broader ecological preservation efforts.

Additionally, a 5G-A network with 10 base stations covers 330 square kilometers of the bay area, facilitating continuous monitoring of key conservation zones. This integrated system allows for real-time tracking of vessel movements and rapid AI-generated alerts for potential violations, leading to investigations of 12 vessels threatening the dolphin population.

As research capabilities expand alongside AI applications, further insights into the behaviors and challenges of Xiamen Bay's Chinese white dolphins emerge, supporting ongoing conservation efforts.