A turning point in human-elephant conflict: Cloud + data + AI to protect our shared land

01 September 2020


China's Yunnan Xishuangbanna National Nature Reserve is located in the southern part of Yunnan province and covers a total area of 242,500 hectares. The reserve is the largest area of primary tropical forest in China and is home to the largest population of Asian elephants in the country.

The Asian elephant is categorized as a first-class protected wildlife species in China and an endangered species on the International Union for Conservation of Nature's Red List. Asian elephants in the country mainly inhabit the rainforest of Xishuangbanna. Thanks to enhanced protection of the rainforest and species population, the number of Asian elephants has increased from around 170 in the 1980s to over 300. However, this increase in numbers has led to an increase in activities and constant need for more land to roam with frequent “incidents” in villages and cities, making human-elephant conflict a frequent occurrence. 24-hour observation and early warnings for such incidents cannot be guaranteed when the elephants are tracked manually by forest rangers or with drones. Furthermore, these conventional methods often generate unreliable results during foggy and rainy days, and could put monitoring personnels safety at risk.


Asian Elephant in wild


Faced with the problem of inaccuracy and delays of such methods of protecting the Asian elephant ecosystem, the management and conservation department of the nature reserve worked with Inspur to build an integrated device, edge, and cloud solution that uses AI, cloud computing, big data, and other advanced technologies. The solution consists of a network of infrared and regular cameras, drones, and other data collection devices throughout the rainforest to capture image and video data in all-weather and in real-time. A high-precision AI recognition model for the species deployed at the edge of the network not only accurately identifies images and video streams received in real time within milliseconds to provide early warnings in a matter of seconds to prevent human-elephant conflict but also synchronizes data that has been cleaned at the edge to the cloud center for training and optimization of the AI recognition model and data aggregation and mining.


Image of an Asian elephant captured by an infrared camera at night

So far, the ecological protection system has sounded more than 2,900 warnings, effectively preventing human-elephant conflict. Meanwhile, the system has utilized the 260,000 images collected to overcome difficulties in recognizing blurred night images, incomplete images of Asian elephants, and other technical hurdles. This has raised AI recognition accuracy for the species from an international average of 60% to 96%, with performance still improving.


Real-time monitoring and recognition of elephants with AI


The ecological protection system was built not only to ease human-elephant conflict and protect Asian elephants, but also to set a new example for the conservation of endangered species and ecosystems.

Supported by technologies such as AI, cloud computing and big data, Asian elephant researchers no longer need to trek through forests and mountains with anticipation and fear, to find footprints, excrement, and food leftovers of the endangered species. By logging into the data aggregation platform of the ecological protection system, researchers can conduct in-depth research and analysis on the numbers, living conditions, upstream and downstream food chains, and distribution and migration patterns of the herds, as well as rainforests humidity and temperature. The platform also provides richer and three-dimensional data sets for deeper research on species conservation and biodiversity.


Technology is one way for humans to explore nature; our human touch is what will ultimately change our world. Making the world a better place with technology is the responsibility and mission of every science and technology creator. The Asian elephant ecological protection system is part of Inspur's efforts to help in conservation of endangered species and ecosystems with technology, and to put this mission into practice. Following the success of the project, Inspur hopes to apply the power of technology to the conservation of primates, birds, insects, and other rare and endangered species, so that technology can make discoveries, data can speak for itself, and we can use technology to protect our shared land.