Unearthing the Secrets of the Ocean Floor

However, with the average oceanic depth of approximately 13,000 feet (4,000 meters), Ved devised a method to dive even deeper- MiDAR. The challenge lies in light, as the further one descends, the lesser the sunlight available to illuminate objects in a manner visible to humans. Nevertheless, fish have adapted to perceive a far wider spectrum of light than humans do, and Ved utilized this concept in MiDAR. High-intensity lasers emitting blue and ultraviolet light, wavelengths that travel particularly well underwater, enable drone-mounted MiDAR devices to capture details down to about 1,650 feet (500 meters). By combining MiDAR imaging with a submersible, it could reach nearly any depth. MiDAR also represents the first technology to employ light for data transmission back to the base—a ground-breaking advancement allowing MiDAR-equipped robots to chart the ocean floor systematically without frequent resurfacing.

This intricate 3D imaging produces enormous volumes of data—hundreds of terabytes for each coral reef. Analyzing this data effectively demands immense computational strength, prompting Ved to create NeMO-Net. The most recent high-performance processors empower AI and ML to swiftly and accurately convert raw data into maps, from which researchers can derive significant insights—such as the condition of a coral reef. To assist AI in constructing the essential models for understanding this data, NeMO-Net has engaged citizen science and the widespread availability of powerful personal computing by gamifying the process of coral identification. Individuals worldwide access NeMO-Net to color coral and categorize this remarkable environment. Their outcomes, peer-reviewed and input into supercomputers, yield the precise maps needed by scientists and conservationists.

An essential technology that connects all of Ved’s creations is the processor, which has become omnipresent due to computing entities like Arm. Central processing units (CPU) drive his innovations: FluidCam utilizes extremely high-performance chips across various architectures, including field-programmable gate arrays that enhance specific fluid lensing processes. Arm microprocessors operate the crucial circuitry that manages MiDAR, facilitating data capture and processing with efficient and compact chips, performing computations (previously necessitating a supercomputer) on a device carried by a small drone.