AI’s ‘Eye’ Spreads in All Directions… Smart ‘Vision AI’ Expands

[Big Trend] Vision AI technology industry penetrates... Related startups are also growing

SK Telecom installed an unmanned multi-use cup collector equipped with artificial intelligence (AI) video analysis technology at its headquarters in Euljiro, Seoul. If customers pay a deposit at an in-house café, use a multi-use cup, and return it to the collector, they receive their deposit back. SKT first launched the project at Starbucks in Jeju Island in 2021.

Machine vision is the core technology behind this retrieval machine. With the development of AI technologies such as deep learning, machine vision has evolved into vision AI, which can identify objects more quickly and intelligently and detect defects. Vision AI refers to technology that analyzes image signals received from cameras or sensors and extracts information similar to the human eye. Looking at related industries on Wednesday, not only large companies but also startups are releasing various services such as autonomous driving and smart city technologies.

Vision AI technologies include ▲ cameras that recognize images or videos like the human eye ▲ radar ▲ laser-based sensors such as LiDAR, which use light waves, depending on the underlying technology. Advanced camera vision can detect in real time whether a person in a video is wearing safety equipment and send a notification if a problem is detected. LiDAR sensors are not affected by changes in weather or lighting conditions and can accurately detect objects even at night.

Superb AI has established a congestion analysis system at KT Wiz Park in Suwon, home of the KT Wiz professional baseball team. Using around 50 CCTVs in the stadium, the system identifies the density of each area in real time. Audience members can also check congestion levels on large electronic display boards. It is the first time that an AI-based congestion analysis system has been introduced in a Korean professional sports stadium. Superb AI plans to cooperate with KT to expand the service to other stadiums, performances, and exhibition halls.

Vueron Technology, which develops LiDAR systems similar to those used in autonomous vehicles, recently conducted a project in Eunpyeong-gu, Seoul. The company analyzed traffic flows of vehicles, pedestrians, and bicycles in specific sections. The data is expected to be used as a reference for long-term policies to improve transportation and pedestrian environments.

Vision AI is also becoming essential in the recycling ecosystem, such as separating and processing PET plastics from bottled water. Innovus (Unmanned Recover Thusam) and Superbin (Nefron) use cameras embedded in their facilities to check and classify not only the material but also the contamination level of bottles in real time. “The AI screening function identifies the state of resources within 0.8 seconds,” said an Innovus official.

As such, Vision AI technologies are increasingly being used in industrial sites, construction sites, public facilities, and the defense industry. In manufacturing, defects are checked in real time and production efficiency is improved. It is already considered an essential technology for smart cities, for example in measuring congestion at large airports in Korea.

Sapeon has introduced Sapeon Safety, an AI-based security monitoring solution, along with AI solutions specialized by industry. Walnut AI Lab, a startup spun out of Seoul National University’s lab, provides machine vision technologies for inspection equipment used by Korean pharmaceutical companies and others.

The global machine vision market is growing mainly in the United States and Europe. According to market research companies, Cognex of the United States, Teledyne Imaging and Basler of Germany, and Omron of Japan are leading the market, with Japan’s Keyence recognized as a global powerhouse. Applications are also expanding to medical AI diagnostics and 3D recognition.

Large Korean companies are also adopting Vision AI for real-time parcel classification, autonomous driving, and monitoring of in-vehicle conditions. “Vision AI can also act as an ‘eye’ for unmanned vehicles such as drones and humanoids,” an industry source said. “We expect its use to continue expanding in the future.”

https://n.news.naver.com/article/008/0005179064?ref=blog-ko.superb-ai.com