There are various risks and problems caused by human factors lurking in manufacturing sites.
We offer solutions to manufacturing site issues with our easy-to-install sensors that integrate a camera and in-house developed algorithm to detect human bodies, faces, and objects.
- The remote operator starts the machine without noticing that the machine is under maintenance.
- The machine starts operating while someone is still in the prohibited area.
- An unauthorized subordinate operates equipment using his/her boss's ID.
- Tools are often left out or lost after work.
For example...
This sensor accurately detects human bodies (workers) even when they are wearing special clothing that makes it difficult to identify their characteristics or when their postures are varied. It distinguishes between the “machine” and the “human body,” which has been difficult with conventional sensor methods, and detects only the human body.
Detectable with various types of clothing
Compatible with various postures
This sensor identifies specific people registered in advance on the edge side (sensor side). Since online data processing is not required, it can reduce information security risks compared to cloud processing methods.
This sensor detects the difference between the state before work and the state after completion of work, allowing you to immediately notice any missing tools or items that have been left behind.
*Currently under development
Frequently Asked Questions
- Q1. Tell me about adoption results at manufacturing sites.
- - A. We have a proven track record in “facial recognition” for identifying workers who operate manufacturing equipment and “human body detection” for detecting workers and maintenance personnel approaching manufacturing equipment. We cannot disclose specific applications and customer names, but we have received inquiries from many customers.
- Q2. It says that various postures and clothing can be detected, but to what extent is detection possible?
- - A. Postures such as standing, squatting, crouching, and crawling can be detected. We also have solutions that make it possible to detect someone who has fallen. Regarding clothing, detection accuracy will vary depending on the specific pattern and color, as well as the installation environment of the equipment, so if you could let us know the specific conditions, we will consider your request and respond accordingly.
- Q3. Has it already been released as a product? Can I buy it?
- - A. Some facial recognition sensors are already available as general-purpose products (see here). Recommended products will vary depending on your application and specifications. Please contact us and our staff will get back to you with more details. Object detection sensors are not yet available as commercial products and are currently under technical review. If you have any specific requests, we will be happy to provide you with further information.
- Q4. Can I borrow it?
- - A. We will consider lending you a sample after hearing about what you want to achieve and the challenges you face. We will also consider providing engineering support by sending our engineer to your site for consultation.
- Q5. What are OMRON's unique strengths and features that differentiate it from its competitors?
- - A. Our strength lies in our in-house developed algorithm (OKAO Vision) that can operate at high speed even with low-performance CPU functionality in an embedded environment without using the cloud. By utilizing our proprietary image data and AI learning know-how we have accumulated over many years, we have sold more than 1.5 billion licenses (as of July 2025). High detection accuracy under niche conditions is a key feature. Another feature is that, because it is an embedded type (cloudless), it can be offered at a relatively low cost compared to other detection methods, thereby reducing security risks.
Color Sensor to Build into Equipment B5WC
It can detect various color changes with high accuracy.
A new type with a built-in algorithm has been released, making it even easier to use.
Top-Down View Human Body Detection Software AM1
It outputs worker location information and detects the time spent at each process.
This contributes to improving productivity efficiency (reducing takt time).
It can also be operated with a commercially available web camera and a personal computer.
Contact Us
Contact Us
Related Content
“OKAO” and “OKAO Vision” are trademarks or registered trademarks of OMRON Corporation in Japan and other countries.
Safety Precautions
- The sensors described in this document do not comply with various safety standards. When using them for purposes such as hazard control, use them as additional sensors after taking safety measures stipulated by laws and regulations.
- Image recognition technology is used. Due to its characteristics, there is a possibility of non-detection or false detection, so please be sure to check the applicability and performance in the actual usage environment before use.
