"AI Technology for Tracking and Identifying Great White Sharks"

2019-10-14

The computing power has been continuously enhanced in the past few years. The combination of big data and improved algorithms has enabled artificial intelligence to move out of research laboratories and become the technological core of many innovative applications. For example, robot surgery, hurricane forecasting systems in the Caribbean, personalized classroom teaching - providing children with the best education, and predicting the best sales leads. Recently, some members of the Salesforce Einstein team collaborated with the Benioff Ocean Initiative at the University of California, Santa Barbara to use artificial intelligence near the Southern California coast to visually identify the presence of great white sharks. They hope that the application of artificial intelligence can help keep beaches safer, encourage more informed public, and help people learn how to share the waves with wildlife. They have designed a new method that utilizes artificial intelligence to identify the presence of sharks in coastal waters. The Salesforce research team believes that they can train the Einstein visual algorithm, which is typically used for visual search, brand detection, and product recognition, to process drone videos and identify the unique attributes of sharks. They collaborated with the team led by McCauley to design an innovative workflow that does not require water contact for video collection and processing. The first step is to collect a large amount of video data along the coast using drone technology. McCauley and his team worked with UCSB students to fly drones away from the beach and capture a 10-mile stretch of coastline. These video data will serve as the foundation for building the artificial intelligence model. In the second step, the Salesforce Einstein team collaborated with the marine biologists team at UCSB to train the algorithm on a series of factors. For example, it must learn to distinguish between shark species - such as leopard sharks and great white sharks, and it must exclude non-shark objects - such as floating kelp that may resemble sharks. Additionally, it must determine whether it is seeing a new shark or a shark it has seen before. Through repeated experiments, the algorithm becomes smarter. Training an artificial intelligence is a complex process that requires a lot of effort. The research team did not develop this shark recognition artificial intelligence on a whim. In simple terms, the new generation of tools and technologies can help people better assess beach conditions and ensure the safety of visitors. Looking deeper, perhaps in the near future, artificial intelligence autonomous drones installed on every lighthouse can assist lifeguards. These drones can perform regular flights, transmit data back to the tower, and participate in maritime rescue operations, not just limited to identifying and tracking great white sharks. Some people say that this is a way to realize dreams through reality. "Artificial intelligence is doing amazing and bold things, inspiring customers to take this leap. But when it comes to practical involvement, many people are reluctant to truly dream. Starting with practical applications can help drive the practical impact of artificial intelligence on the world."

RECENT AI NEWS

RECENT AI TOOLS