The Dong-A Ilbo IT Science Team introduces noteworthy technologies, trends, and companies in IT, science, space, and bio fields. "What is this company?" The behind-the-scenes stories of tech companies changing the world with technology! From ideas that surprised the world to the current concerns of founders, they delve into everything you were curious about.
A screen captured by KLE's 3D machine vision technology with an accuracy of 0.1mm at 4 million points on a human face. Provided by KLE
Robots are now an inevitable future. The era where robots assembling cars in factories, folding laundry at home, assisting surgeries in hospitals, and patrolling dangerous battlefields is no longer surprising.
So, how accurate are the 'eyes' of these robots? To completely replace humans, they must assemble car parts in the correct position and aim at surgical sites without error. A company creating these 'eyes' or 'visual intelligence' for robots was explored. In the global market, China's Mech-Mind Robotics, Germany's Isra Vision, and Norway's Zivid are considered leading players. In Korea, 'KLE,' founded by Lee Jin-han, a former research professor at Hanyang University's Department of Convergence Electronics Engineering, is making significant strides since its establishment in 2021.
"Value humans, leave dangerous tasks to robots"
This is the core motto of 'KLE,' which creates the 'eyes' of robots. 3D machine vision refers to the technology that enables robots to perceive and operate in environments like humans. KLE's technology, which captures 4 million points within a scene with a 0.1mm error in just 1.9 seconds, targets Korean and Japanese automobile manufacturers.
KLE's core product, the 3D machine vision camera COPIC 3D series, implements high-precision 3D measurement performance with an error level of up to 0.1mm. Based on NVIDIA's parallel processing architecture, it secures real-time computation capabilities, enabling fully unmanned automation of precise assembly and inspection processes that were impossible with traditional 2D machine vision. The proprietary technology that extracts precise 3D data without expensive 3D sensors is highly regarded for its industrial innovation.
Recognized for such technological prowess, KLE succeeded in supplying to Hyundai Motor Group's Meta Plant in Georgia, USA, the new electric vehicle plant in Ulsan, and Kia's plants in Mexico and Slovakia. Currently, it is conducting technology demonstration tests with Japanese automakers such as Honda, Toyota, and Mitsubishi Motors. The following is a Q&A with Lee Jin-han, CEO of KLE.
KLE's 3D machine vision cameras installed in overseas automobile factories. Provided by KLE
Why is advanced 3D machine vision technology necessary in automotive and other manufacturing plants?
Existing 2D vision technology is limited to planar information, clearly restricting tasks requiring complex shapes or depth perception. 3D recognition is essential for automating high-difficulty tasks such as assembly processes, welding, dimensional inspection, and defect detection.
In past factories, robots placed parts in similar positions, and humans performed precise tasks. However, to achieve full automation with robots, ensuring 'correct positioning' is crucial. Just like manual work, vehicle glass must be attached accurately, with minimal error. Our 3D machine vision technology allows for accurate positioning, enabling robots to attach glass to car bodies precisely like humans. For tires, even a 0.2mm error can prevent fitting, but our technology meets this condition with a 0.1mm error.
Can it completely replace the human hand?
The human hand has tactile sense, feeling resistance or repulsion. Implementing this in robots is important, but there is a long way to go. Due to the difficulty of this method, 3D machine vision technology is used to accurately capture and precisely measure the required position for assembly work.
Performance tests are underway with major Japanese automakers. What is the core strategy for penetrating the Japanese market?
Our technology has been evaluated as overwhelmingly precise compared to competing products. Japanese customers are characterized by verifying each specification we propose. As a result of direct verification by Japanese customers using specialized equipment, our product showed an error rate of 0.04mm, while a Chinese competitor showed 0.4mm. Our product demonstrated 10 times superior performance.
What potential have you discovered in the Japanese market?
Japan is a technological powerhouse but a latecomer in factory automation. The Japanese automotive market is three times larger than Korea's, but Korea's factory automation rate is six times higher than Japan's. For us, who create the 'eyes' of robots, Japan is an 18 times larger market than Korea, even with simple calculations.
Japan's culture of valuing craftsmanship has delayed the adoption of unmanned automation. The manufacturing culture in Japan is inherently the domain of artisans. There is a strong perception that the best experts assemble cars. However, facing fierce competition from China, they must consider price competitiveness in the global market. If labor costs continue to rise, they will lose competitiveness. Japanese automakers' 3D machine vision technology is not advanced, leading them to test with Korea's KLE and China's Mech-Mind. This presents a tremendous opportunity for us. Additionally, the aging population and shortage of manufacturing workforce are factors that inevitably accelerate the introduction of robots in factories.
Lee Jin-han, CEO of KLE. Provided by KLE
Can you overcome China's overwhelming economies of scale as a robotics powerhouse?
In fact, Korea was more earnest about 'factory automation' before China. Korea started using robots extensively in manufacturing plants before China. That data has accumulated as a tremendous asset in Korea. With the world's highest level of factory automation systems, Korea possesses high-quality manufacturing data accumulated by major manufacturing facilities in the automotive, semiconductor, steel, and shipbuilding industries. It is the forefront of implementing Physical AI.
Attention should be paid to the remarks made by Hyundai Motor Group Chairman Chung Eui-sun when President Lee Jae-myung met with NVIDIA CEO Jensen Huang on the 31st of last month. (Chairman Chung said at the time, "Hyundai Motor Group possesses high-quality data through various mobility solutions, parts, and complete vehicle manufacturing ecosystems. It is the optimal company to provide future mobility solutions based on AI, as well as develop Physical AI.")
What areas will you challenge next?
Robots are being deployed in high-risk areas for price competitiveness, high efficiency, and safety. The next area after automobiles is logistics. In automotive manufacturing plants, standardized parts with fixed specifications and frameworks are handled.
However, the logistics sector is a world of irregularities. Consider a Coupang logistics center. Items from Coupang are packaged in plastic. Robots cannot easily stack items wrapped in irregular plastic in the delivery order like humans do. These irregular tasks are areas to be challenged in the future.
ⓒ dongA.com. All rights reserved. Reproduction, redistribution, or use for AI training prohibited.
Popular News
