Research Report

8K Combined with AI: Towards Innovative Disaster Coverage

Published: September 1, 2019

Is artificial intelligence (AI) capable of detecting human figures from an 8K ultra-high-definition (UHD) image with a resolution of 33 million pixels? With a resolution 16 times that of HD (2K), 8K broadcasts started in December 2018 in Japan, which enabled UHD TV programs to be delivered to the viewers. Meanwhile, efforts to employ 8K for disaster risk reduction activities are also underway, such as detecting earthquake faults and other disaster factors as well as finding people seeking rescue, by regarding 8K as a “visual sensor for protecting people’s lives” utilizing its UHD imaging technology.1) In particular, aerial filming from a helicopter can extensively survey disaster-affected areas. If affected sites and persons can be instantly detected from an enormous amount of 8K image data, it may lead to further sophistication of disaster coverage. To achieve this, we assumed that incorporating 8K with AI’s image recognition technology would be effective and tested the hypothesis in cooperation with NHK Science & Technology Research Laboratories (STRL). This paper reports the results of the test and discusses the potential of 8K broadcasts for disaster risk reduction by reviewing the ongoing efforts of combining 8K, AI, and 5G in other fields such as medical care, security, and infrastructure maintenance.

The NHK Monthly Report on Broadcast Research

Masaru Yamaguchi

in Japanese