Shimizu Corporation has developed a system that uses AI to streamline construction management for mountain tunnel construction. AI analyzes the work situation inside the tunnel from live video from the web camera and automatically identifies work content near the face. Judgment results can be used to share information among construction employees and automatically create job profiles through chat tools and cloud storage. As a result of implementing real-site testing, we have confirmed that site staff work and waiting times in the mine are reduced compared to conventional methods.
Outline of automatic judgment system according to AI cycle (from press release document)
It was announced on the 20th as "AI cycle automatic judgment system". The system includes devices such as network cameras, high-speed communication networks, analytics computers (PCs), cloud storage, and chat engines.
According to the company, at conventional mountain tunnel construction sites, it is difficult to grasp the work status inside the tunnel from outside the tunnel, and the work process is adjusted by making inferences based on the information. limitations from camera images as well as the experience and intuition of workers. site staff. If these assumptions are not met, employees entering the mine will be forced to wait unnecessarily.
The system uses PC image analysis AI to analyze live video from network cameras installed behind the work surface and automatically determine the nature of work in the vicinity of the work surface. AI-based judgment results will be transferred to cloud storage and used to share information with construction staff and automatically create job profiles through chat tools.
So far, the company has tested the system at two construction sites (Chuo Shin-Kobutsu Expressway Tunnel and Toga Tunnel (section 1)) and was able to reduce on-site staff waiting time by about 40%. underground. During the construction of the Shinkobutsu tunnel, we also used the ``Mountain Tunnel Construction Data Platform'', which aggregates and centrally manages construction data accumulated daily. In a trial targeting rock face recognition, automating accounting records reduced staff work time by about 80%.
In the future, the company will use data accumulated on the platform to detect abnormalities early and disseminate more information on-site to improve safety.
Invite partners to see the activities of Pacific Group Co., Ltd.
FanPage: https://www.facebook.com/Pacific-Group
YouTube: https://www.youtube.com/@PacificGroupCoLt
