Ridge-i Ltd. (located in Tokyo Chiyoda-ku, CEO: Yanagihara Takashi, thereafter referred as Ridge-i), company specialized in AI and Deep Learning consulting and development, and Ebara Environmental Plant Co., Ltd. (located Tokyo Oda-ku, CEO: Watanabe Yoshio, thereafter referred as Ebara Environmental Plant) have started a joint operation by developing an AI aiming at replacing a skilled crane operator's eye to perform garbage separation. After a successful development and experiment phase, the model was finally deployed to production.
By watching live camera footage of a pit containing a huge variety of garbage, the system is able to analyze the scene, take operating decisions and operate the crane remotely to process it, without requiring any human action - all of it being performed in difficult situations involving garbage stirring, ... During the development process, replacing the operator's eye was a huge challenge, but by leveraging on latest advances in Deep Learning an AI able to differentiate various types of garbage was successfully delivered (figure 1).
figure 1: Illustration of a garbage incinerator
Inside the incinerator, it is essential to stabilize several factors such as exhaust gas generation or the combustion status. To achieve this, a crane is needed to stir the garbages in the pit and to evacuate special garbages.
In the present setting, a human operator visually checks the pit state, and manually operates the crane to stabilize the combustion status when needed. However, due to the discrepancy of individual skills and the limited amount of human ressources to be expected in the future, the company has heavily invested to make the whole process automated.
Therefore, Ridge-i and Ebara Environmental Plant started a collaboration in order to design an AI powered system able to autonomously analyse real-time camera feed of the pit, take operating decision and control the crane.
By leveraging on the latest state of the art AI techniques, the resulting system is now able to analyze the entire scene pixel-wise, detect torn plastic bags and check the complex nature of the garbage within them. Moreover, two up-to-date unsolved problems encountered in previous attempts, namely "accurately analyze the inside of the furnace without disturbance from the chaotic environment" and "deal with special garbages - which can negatively affect the equipment integrity and combustion status if thrown in large quantities" have been successfully solved. (figure 2)
Figure 2: Appearance of a garbage pit and output image from the AI classifier.
The newly released system required extensive knowledge from both companies: Ebara Environmental Plant spent years to gather a complete understanding about the various challenges they faced in their daily operation. Then Ridge-i shared their expertise in computer vision based on state of the art deep learning technologies, and after receiving a large amount of feedback through the entire extent of this Agile-managed project, we finally developed this new technology able to compete with the most experienced operator's eye.
From now on, Ridge-i and Ebara Environmental Plant will implement this new AI-powered system to both existing garbage incinerators and the one which will be built in the future, therefore we are very likely to see AI continuing to take control of new facilities in the future.
From now on, Ridge-i will continue to focus on bringing state of the art AI solution to the business world, and strives to contribute to modernize industries in numerous various fields in the future.
▼ About Ridge-i
Ridge-i leverages on the latest science advances to provide state of the art AI solutions to fit and optimize business needs.
Placing ourselves at the boundary between AI consulting and development while being aware of the various business word constraints, we aim at delivering effective and persistent state of the art solutions to our clients.
Ridge-i Co., Ltd.
Tokyo Chiyoda-ku, Otemachi 1-6-1, Otemachi Bldg 442
CEO: Yanagihara Takashi