Ko-Mapper Collaborates with HD Hyundai Heavy Industries on Steel Structure Diagnostics Using AI-Based Facility Safety Inspection Solution

사진 확대

Limitations of inspecting aging infrastructure. Around the world, aging infrastructure is emerging as a quiet disaster. In South Korea as well, the share of bridges, tunnels, dams and other facilities built during the industrialization era that are now more than 40 years old is rising sharply. The traditional method, in which skilled inspectors rely on ropes to visually check for cracks and record findings by hand, has clear limits in terms of safety risks and the ability to secure objective data.

Ko-Mapper's AI facility inspection technology. Founded in 2019, Ko-Mapper is drawing attention in the industry with its AI-based facility safety inspection solution. Led by CEO Kim Dal-ju, who studied civil engineering and researched satellite image processing, the startup is changing the paradigm of infrastructure management through Physical AI technology. The company name, 'Ko-Mapper,' is a blend of 'Korea' and 'Mapper,' reflecting its goal of precisely mapping domestic facilities and infrastructure and transparently revealing hidden risks as data.

Core performance of KO-DETECT PRO. Ko-Mapper's flagship product is KO-DETECT PRO, an AI-based damage detection and analysis solution. The AI analyzes image data collected by drones, robots and sensors to automatically detect a wide range of facility defects, including cracks, delamination, efflorescence and exposed rebar. In the past, people had to inspect tens of thousands of photos one by one, but Ko-Mapper's AI quantifies defects with a consistent standard and without omissions, then automatically generates reports and provides them in database form.

Technical differentiation from competing solutions. Ko-Mapper's biggest technical advantage lies in speed and field optimization. Most global commercial software takes more than three hours to process around 100 photos when modeling 2D images into 3D. Ko-Mapper focused on the fact that heavy 3D modeling is not necessary for facility inspections, and developed its own image stitching technology, cutting a task that once took three hours down to just two minutes. Image analysis processing time has been reduced from four minutes to around 10 seconds, and unofficially the company has achieved a record speed of 7.46 seconds.

Technology optimized for field use. Ko-Mapper has lightened its AI engine so it can run on ordinary office PCs or laptops without expensive GPU equipment. This makes immediate analysis possible even in public institutions or defense-related sites where cloud use is restricted for security reasons. Its core strength is not just an AI that detects well, but an AI that actually works in the field.

A demonstration case in the Japanese market. Even in the conservative civil engineering market, Ko-Mapper's technology is being adopted quickly as labor shortages and aging populations become more serious. A recent demonstration in Japan proved its efficiency. A large water tank facility, as tall as a 10-story apartment building and requiring a year of inspection by local experts, was completed in just 7.5 hours using commercial drones and Ko-Mapper's AI solution. Japanese counterparts responded with an immediate contract.

Domestic adoption status. In South Korea as well, adoption is moving into real-world use, centered on public institutions and major construction companies. In one case, experts in the field identified 97 cracks through visual inspection, while Ko-Mapper's AI found 210. The example is seen as evidence that technology can fundamentally block the risk of poor inspections caused by human subjectivity.

Results of collaboration with HD Hyundai Heavy Industries. Through the Seoul Center for Creative Economy & Innovation's Deep Tech Value-Up Program 2025, Ko-Mapper secured a practical collaboration opportunity with HHI. Ko-Mapper, which had already shown strength in diagnosing concrete structures, worked with Hyundai Heavy Industries to jointly build a safety inspection model for massive steel structures such as Goliath Cranes. Inspecting a crane more than 100 meters tall while relying on ropes is extremely dangerous.

Expansion of steel structure diagnostics technology. At HHI's site, Ko-Mapper newly configured a dedicated AI engine to detect the fastening condition of rivet bolts, signs of corrosion and microscopic cracks in welded joints. By handling real data from a complex and tightly secured heavy-industry site, the company demonstrated both technical performance and field applicability. Rather than stopping at a one-time PoC, it held ongoing meetings with HHI and refined the model to a level that can be deployed immediately on site.

Plans for overseas expansion. For its overseas push, Ko-Mapper is prioritizing Japan and the US. Japan has already provided an export reference, while in the US the company established a local joint venture last September and completed federal government procurement registration. In March, it is set to sign formal contracts with the Panama Canal Authority (ACP) and the Panama Railway Authority for facility diagnostics. Panama has praised Ko-Mapper's objective data analysis capabilities.

Expansion of business areas. The company is broadening its business beyond concrete and steel, preparing to enter agricultural inspection, defense and a new SAR-based image analysis field that can observe the Earth's surface even from above the clouds. CEO Kim Dal-ju explained, "Facility safety does not reveal its value until an accident happens, but by then it is already too late." He added, "Our goal is to use technology to accurately identify even the subtle risks that people tend to miss before an accident occurs."

Q1. What are the features of the steel structure safety inspection solution that Ko-Mapper developed in collaboration with HD Hyundai Heavy Industries?

Ko-Mapper worked with HHI to jointly build a safety inspection model for massive steel structures such as Goliath Cranes. It newly configured a dedicated AI engine that detects the fastening condition of rivet bolts, signs of corrosion and microscopic cracks in welded joints, and refined the model to a level that can be deployed immediately on site.

Q2. Which company achieved faster speeds than conventional methods in AI-based facility safety inspections?

Ko-Mapper developed its own image stitching technology and reduced a task that once took three hours to just two minutes. Image analysis processing time has also been cut from four minutes to around 10 seconds, and unofficially the company has achieved a record speed of 7.46 seconds.

Q3. What is Physical AI?

Physical AI refers to AI technology that understands the language of the field. Through Physical AI, Ko-Mapper is changing the paradigm of infrastructure management by having AI analyze image data collected by drones, robots and sensors to automatically detect facility defects such as cracks, delamination, efflorescence and exposed rebar.

This content was prepared by MaeKyung Biz Content Planning Team based on official announcements and press releases.