October 7, 2024
EXEO Digital Solutions, Inc. (head office: Minato-ku, Tokyo; Koichi Maeda, President; hereinafter "EDS"), a core subsidiary of the EXEO Group, Inc. (head office: Shibuya-ku, Tokyo; Tetsuya Funabashi, President; hereinafter "EXEO Group") that develops system solutions, announces the launch of a new project to support quality inspections of gas appliances installed by Osaka Gas Marketing Co., Ltd. (head office: Chuo-ku, Osaka; Takeshi Morisaki, President; hereinafter "Osaka Gas Marketing") using an AI image analysis system that was developed by SecondXight Analytica, Inc. (head office: Chiyoda-ku, Tokyo; Hirokazu Takayama, President; hereinafter "SecondXight Analytica").
１．Project Background and Purpose
When installing gas appliances, a double-check is executed after installation. The person who installed the appliance performs a visual check, and a supervisor performs another check using installation photographs. However, with more than 100,000 installations each year, finding a system or mechanism to ensure quality precision while also allowing for checks to be performed quickly has become a challenge.
This project implements an AI image analysis system in response to this challenge. The system analyzes the installation on-site in real-time from photos taken during installation with an advanced, high level of accuracy. This reduces the number of personnel required to perform checks and improves efficiency. Through the use of AI, the massive undertaking of checking installations is performed quickly and homogeneously, and it aims to improve productivity while ensuring safety, both of which are matters of great importance to Osaka Gas Marketing.
2. Technology Used and Results
EDS and SecondXight Analytica offer installation quality feedback in real-time for nine items for inspection, utilizing deep learning* technology. This technology enables high-resolution analysis at the pixel level for complex inspections, including checking the tightness of pipe connection screws and sealant application. This technology has already been introduced in the communications installation field by EDS and is helping to improve installation quality.
3. Project Progress
This project has advanced to development after PoC (proof of concept) experiments
News Release
in February 2024, and it is expected to be released in April 2025.
EDS will steadily advance this project to improve the efficiency of installation quality inspections for Osaka Gas Marketing, helping to solve future labor shortages through the use of AI and various other IT technologies.
* Deep learning is a field of machine learning. The technology learns patterns from a large amount of data, enabling the prediction and classification of new data. More complex and abstract characteristics can be extracted from data by using multilayer neural networks, and this technology has proven itself in facial recognition and a variety of other fields.
【Contact for inquiries】
Public Relations Team, Corporate Communication Unit, EXEO Group, Inc.
TEL. +81-3-5778-1075 E-mail：contact@en2.exeo.co.jp
