Keisoku Kensa and Mahoroba Soken to Co-Develop AI for Non-Destructive Pipe Deterioration Assessment from a Single Photo

Mahoroba Soken Co., Ltd. (Representative Director: Kazuki Komori, Head Office: Nara City, Nara Prefecture, hereinafter "Mahoroba Soken") and Keisoku Kensa Co., Ltd. (Representative Director: Toshihiro Sakamoto, Head Office: Kitakyushu City, Fukuoka Prefecture, hereinafter "Keisoku Kensa") will begin joint development in June 2026 of AI software that enables non-destructive assessment of plant pipe deterioration from a single photograph.

By combining Keisoku Kensa’s inspection field knowledge accumulated over more than 50 years with Mahoroba Soken’s AI development expertise specialized in predictive maintenance for industrial equipment, the companies aim to launch the product by fiscal 2027.

Keisoku Kensa and Mahoroba Soken to Co-Develop AI for Non-Destructive Pipe Deterioration Assessment from a Single Photo

### ■ Background and Challenges: Aging Plants and Maintenance Staff Shortages

Many industrial plants in Japan, constructed during the period of high economic growth, are still in operation, making aging infrastructure a pressing issue. According to documents from the Ministry of Economy, Trade and Industry (METI), aging equipment increases the risk of accidents and can lead to equipment malfunctions and unstable product quality.

At the same time, maintenance personnel responsible for inspecting these facilities are aging and facing severe labor shortages, making it difficult to pass on inspection know-how that has traditionally relied on veteran technicians’ intuition and experience.

Conventional inspection methods have primarily followed a time-based maintenance approach—replacing parts when a certain period has elapsed—"fix it when the time comes." Due to the high level of expertise required, inspections have depended heavily on a limited number of skilled professionals, creating significant challenges.

### ■ Overview of the Pipe Inspection AI: Preliminary Diagnosis by Anyone, from Just One Photo

The software being jointly developed by both companies will use photos of pipes to enable AI to assess the degree of deterioration and determine inspection priority.

The goal is to diagnose pipe conditions non-destructively—without causing any damage—using just a single image captured by a smartphone or similar device. The system will also support images taken by drones, enabling remote inspection of pipes in hard-to-reach locations such as high elevations or confined spaces, without requiring personnel to enter hazardous areas.

### ■ Strengths of Both Companies: Field Expertise × Predictive Maintenance-Focused Software Development

A key feature of this development is leveraging Keisoku Kensa’s extensive field knowledge gained through over 50 years of inspection operations.

Rather than relying on widely used general-purpose generative AI, the companies aim to achieve high, practical diagnostic accuracy by developing specialized software focused on "predictive maintenance"—predicting equipment anomalies and linking them to maintenance actions. Mahoroba Soken specializes in AI and software development for predictive maintenance, and the collaboration brings together both companies’ technical capabilities and expertise.

### ■ Comment from Kazuki Komori, Representative Director, Mahoroba Soken Co., Ltd.

Kazuki Komori, Representative Director, Mahoroba Soken Co., Ltd.

In plant maintenance, a major challenge is how to pass on inspection techniques, long supported by the experience of skilled technicians, to the next generation.

We will develop dedicated AI deeply tailored to the domain of pipe inspection, along with software that emphasizes usability and intuitive interface design—ensuring field personnel feel it is truly practical when they use it.

For our mission of realizing predictive maintenance through software, our partnership with Keisoku Kensa, which has long supported inspection operations, is invaluable.

By cherishing this partnership, we aim to create a new standard for plant safety—where anyone can detect early signs of deterioration from just one photo.

### ■ Comment from Toshihiro Sakamoto, Representative Director, Keisoku Kensa Co., Ltd.

Toshihiro Sakamoto, Representative Director, Keisoku Kensa Co., Ltd.

Since our founding over 50 years ago, we have supported the maintenance and management of social infrastructure and industrial facilities, contributing to people’s safety and peace of mind.

The field knowledge and technical expertise we have cultivated over the years represent one of our greatest corporate assets.

However, in recent years, with the increasing aging of equipment and shortage of technical personnel, maintenance and management operations are being asked to achieve higher levels of efficiency and sophistication than ever before.

This joint development project represents a bold initiative to fuse our long-standing maintenance expertise with Mahoroba Soken’s advanced AI technology, evolving into a system that allows anyone to leverage the experience and judgment of skilled technicians.

To realize our corporate philosophy of "nurturing those who protect, and continuing to safeguard safety and security," we hope this software will contribute to preventing equipment failures before they occur, improving maintenance efficiency, and helping build a society where predictive maintenance becomes the norm.

Our mission extends beyond protecting equipment—we are committed to safeguarding the safety and peace of mind of the people who use these facilities.

Going forward, we will continue to create new value by integrating AI with maintenance technologies, always staying close to the field.

### ■ Company Overview

＜Overview of Mahoroba Soken Co., Ltd.＞

Company Name

Mahoroba Soken Co., Ltd.

Representative Director

Kazuki Komori

Date of Establishment

February 25, 2025

Head Office Address

3-1 Hashimoto-cho, Nara City, Nara Prefecture 630-8217, Japan

Business Activities

Development of AI and software specialized in predictive maintenance for equipment anomalies

Website

https://maholab.co.jp

＜Overview of Keisoku Kensa Co., Ltd.＞

Company Name

Keisoku Kensa Co., Ltd.

Representative Director

Toshihiro Sakamoto

Date of Establishment

November 1, 1974

Head Office Address

1-8-3 Jinbata, Yahata-Nishi Ward, Kitakyushu City, Fukuoka Prefecture 807-0821, Japan

Business Activities

Non-destructive testing (maintenance inspections for chemical plants, welding inspections for steel structures), material evaluation, structural analysis, various measurements (vibration, civil engineering, stress, environmental), etc.

Website

https://www.keisokukensa.co.jp