Demonstration test of a technology that enables image recognition AI to estimate the corrosion depth of steel materials for the digital transformation of social infrastructure inspection

Combining drones and image recognition AI to improve the efficiency of road bridge inspection and reduce maintenance costs

News Highlights:

TOKYO - October 3, 2024 - NTT Corporation (Headquarters: Chiyoda Ward, Tokyo; Representative Member of the Board and President: Akira Shimada; hereinafter "NTT") and NTT e-Drone Technology Corporation (Headquarters: Asaka City, Saitama Prefecture; President: Masahiro Takizawa; hereinafter "NTT e-Drone Tech") started demonstration tests of corrosion inspection of steel materials using drones and image recognition AI on a road bridge in Kumagaya City (Headquarters: Kumagaya City, Saitama Prefecture, Mayor of Kumagaya: Tetsuya Kobayashi) on Monday, September 2, 2024.
　In this experiment, we used AI to detect the corrosion of steel from images of road bridges taken by drones for the digital transformation of road bridge inspection and verify for the practical application of an inspection method that automatically estimates the depth of corrosion (the amount of damage in the steel section due to corrosion.) As a result, the steel thickness in the corroded area, which is currently measured by inspectors using ultrasonic equipment, can be replaced by drones taking images, thereby reducing maintenance costs through more efficient inspection work.

1. Background

Road bridges are important infrastructure facilities that support our economy and life, but the deterioration of these facilities has become a major social problem. Corrosion of steel materials is one factor causing road bridge deterioration. Corrosion that occurs in the steel material causes the loss of cross-section of the steel material as it progresses, and the durability and load-bearing performance of the equipment gradually deteriorates, which may eventually lead to failure or collapse. Therefore, the facility management needs to understand the thickness of the steel where corrosion occurs.
　However, with the current inspection method, it is difficult to determine the thickness of steel at the corroded area. Currently, inspectors visually check the appearance of the equipment for corrosion, so it is impossible to determine the depth of corrosion (the amount of damage in the steel section due to corrosion.) Although there is a method to measure the thickness of steel materials using ultrasonic waves at places where corrosion is severe, it is not easy to do this because it requires a probe to be placed at the measurement point and requires a large amount of operational expense for the entire facility. Also, when inspecting a large road bridge, costs such as scaffold installation may incur. Against this background, the Ministry of Land, Infrastructure, Transport and Tourism is promoting the introduction of inspection support technology¹. To conduct efficient inspection of road bridges, it is a principle to use the technology in the "Inspection support technology performance catalog" specified by the Ministry of Land, Infrastructure, Transport and Tourism for national highways under its direct control.
　NTT and NTT e-Drone Tech are studying a method to improve the efficiency of inspection of steel structures using drones and image recognition AI as an inspection support technology for road bridges. This method uses image recognition AI to detect corrosion from images of a road bridge taken by a drone, and automatically estimates the corrosion depth (the amount of loss in the steel section due to corrosion) at the location. In addition to reducing the cost of installing scaffolding, which is required when inspecting a large road bridge, the use of image recognition AI makes it easy and inexpensive to identify corroded parts of equipment and measure the thickness of steel materials.
　On September 2, 2024, NTT and NTT e-Drone Tech collaborated with Kumagaya City to test the feasibility of facility inspection using drones and image recognition AI using a road bridge owned by Kumagaya City.

2. Outline of the experiment

The drone captures images of road bridges and uses image recognition AI to detect corrosion of steel materials, estimate corrosion depth, and verify work efficiency and technical accuracy.

(1) [Verification items]

The image recognition AI used in this verification is a customized version of the corrosion detection technology and the steel section loss estimation technology² developed by NTT for communication pipelines.

(2) [Experimental period]

From September 2, 2024 to February 28, 2025

(3) [Verification points for practical application]

It is difficult to always maintain a fixed distance between a drone and a road bridge due to the need for advanced operation technology and the difference in the possible flying space depending on the shape of the bridge. As shown in Figure 2, there is a difference in the pixel resolution (mm/pixel)³ of the corroded area between a distant view and a close view of the same corroded area, which may affect the estimation accuracy of the corrosion depth.
　Therefore, we will customize the image recognition AI so that corrosion depth can be estimated with high accuracy even from images with coarse pixel resolution. At the same time, we will clarify the relationship between corrosion depth estimation accuracy and pixel resolution, and aim at the practical application of this inspection method by determining the photographing conditions (photographing distance, photography equipment, etc.) with a drone during operation.

Figure 2 Difference in Pixel Resolution at Different Shooting Distances for the Same Steel Corrosion

3. Role of each company

[NTT]

Accuracy verification of corrosion detection and corrosion depth using image recognition AI, customization of image recognition AI, and measurement of the amount of loss in steel sections at corroded areas using ultrasonic equipment

[NTT e-Drone Tech]

Efficiency verification of inspection work using drones and image recognition AI, photographing of road bridges by drones, examination of photographing conditions (photographing distance, on-board camera, etc.) during drone flight

[Kumagaya City]

Provision of inspection results of demonstration test sites and road bridges

4. Outlook

Based on the results of this demonstration test, we will evaluate its practicability and plan to introduce it as an inspection support technology in FY2025. In addition to road bridges, we will expand our technology to include various infrastructure facilities such as steel towers and guardrails and contribute to realizing a sustainable society by solving problems such as the increase in the maintenance and management costs of social infrastructures.

¹Ministry of Land, Infrastructure, Transport and Tourism press release: "Call for Bridge Inspection Support Technology -Enhancing inspection support technology performance catalog and promoting utilization of new technology-" August 30, 2024, https://www.mlit.go.jp/report/press/road01_hh_001840.html

²NTT news release: "Established a technology to automatically estimate the depth of corrosion on steel materials in infrastructure facilities from images," May 13, 2024, https://group.ntt/en/newsrelease/2024/05/13/240513b.html

³Pixel resolution (mm/pixel): The actual size (mm) of one pixel in a shot image.

About NTT

NTT contributes to a sustainable society through the power of innovation. We are a leading global technology company providing services to consumers and businesses as a mobile operator, infrastructure, networks, applications, and consulting provider. Our offerings include digital business consulting, managed application services, workplace and cloud solutions, data center and edge computing, all supported by our deep global industry expertise. We are over $97B in revenue and 330,000 employees, with $3.6B in annual R&D investments. Our operations span across 80+ countries and regions, allowing us to serve clients in over 190 of them. We serve over 75% of Fortune Global 100 companies, thousands of other enterprise and government clients and millions of consumers.