(Construction) - Japanese construction company Hazama Ando and Kanazawa Institute of Technology have developed the concrete compaction assessment system using artificial intelligence.
The system proposes the assessment method based on deep learning as an alternative to the conventional method based on visual evaluation. A field test using a prototype of the system was conducted at a concrete product plant, and it was confirmed that the system can display in real time the results of the judgment of completion or non-completion of compaction by artificial intelligence.
|Japanese construction company develops AI evaluation system of concrete compaction|
In the system, the AI repeatedly learns the judgment of concrete experts through deep learning, realizing highly accurate judgment close to that of experts. The video camera images of the concrete surface during compaction are transferred to the computer equipped with the judgment program, and the judgment of compaction is made in real time. The result is displayed on the monitor.
The captured images are automatically divided into meshes and judged for each mesh. The color of the frame changes to red when compaction is incomplete and to green when compaction is complete. In the field test, the AI was able to show the results in real time to the on-site workers engaged in the concrete compaction work. This makes it possible to make a judgment on compaction regardless of the skill level of the worker, and to deliver high-quality concrete construction. In the future, the company will make further improvements to the program and aim to apply it to its concrete product plants. (2021/06/22)
The Daily Engineering and Construction News of Japan