Prioritize workplace safety
Safety, always the central concern to the industrial environment, can be enhanced by edge AI. The Occupational Safety and Health Administration (OSHA) requires businesses to provide personal protective equipment (PPE) to protect employees from hazards that could cause injury, where OSHA Regulation 1910.135 states that employers should ensure that each affected employee wears a protective hard hat while working in areas where there is a risk of head injury from falling objects. According to Mckinsey, the U.S PPE market is expected to grow 12.5 percent a year through 2024. It is good to see employees’ health and safety is being prioritized. Enterprises are also seeking ways to build an intelligent system in an easier and integrated way to reduce accidental risk.
Edge devices embedded with AI are capable of monitoring PPE including hard hats compliance in the work environment in real-time and signaling any PPE violations to safety and maintenance. Computer vision combined with machine learning can automate the process of monitoring PPE compliance. Integrating with CCTV systems, building a pipeline autonomously identify employees who are not using or inappropriately using PPE is easier by seamless integration and flexible development.
Automated real-time hardhat-wearing detection
In today’s article, we will share with you how to deploy an automated real-time detection for hardhat-wearing compliance, along with the alert at the workspace, right on the NVIDIA AI embedded system.
- Edge Impulse Studio to upload dataset, acquire custom data, visualize the data, train the machine learning model and validate the inference results.
- Part of the Flickr-Faces-HQ (FFHQ) (under Creative Commons BY 2.0 license) to rebalance the classes in our dataset.
- Edge Impulse Linux SDK and Edge Impulse command line interface (CLI) .
- NVIDIA Jetson platform for model deployment. We use Jetson Nano Module attached on A203 carrier board, but following list will also support:
Hence, ?kudos to Louis Moreau and Mihajlo Raljic from Edge Impulse, we followed their guide trained an embedded Machine Learning model to detect hard hat, and finally deploy it to the Jason Nano. The Jason NX and the Jetson AGX are both supported.
Follow our wiki tutorial, create an account at Edge Impulse, start from model training to final deployment. This project also has been publicly released. Clone the project, go through every step to get a better understanding. You can use it, modify it and integrate it into a complex application.
You can also clone this Hard Hat Detection Github repository for environment setting up and downloading datasets, however, we will more recommend you use Edge Impulse to build a custom dataset using a camera with Jetson Nano or your PC. Therefore, the accuracy will much more match with real scenarios.
Follow the deployment commands, run your real-time detection
The deployed edge box can enable real-time monitoring of hard hats with respect to the work environment and can send alerts in case of any violations.
Step further on particular application
Regarding the particular scenario, we will more recommend you use a public dataset, combining a custom dataset at Edge Impulse studio. Therefore, the accuracy will much more match with real scenarios.
PPE compliance also includes gloves, masks, goggles, etc. Once you finish custom model training, you can also wrap everything into an image, directly deploy the full PPE detection pipeline right at the workplace. Stay tuned with us for more guidance.
Edge Impulse made ML development easier. Combining compact power-efficient AI systems, the process to deploy the edge AI solving specific workplace safety needs becomes faster and more flexible.
However, to reach zero harm, companies must consequently go far beyond their current practices.
About Edge Impulse
Edge Impulse is the leading development platform for machine learning on edge devices, free for developers and trusted by enterprises. Get started today ?
Work with Seeed Ecosystem
Deploying an AI idea can be faster, flexible, even scalable for everyone. Seeed Jetson Platform targets on helping educators, developers and enterprises deploy ML in the real-world. By consolidating Seeed’s best-in-class hardware, cutting-edge technology from our software partners and all developers from the community, we aim at emerging all kinds of AI scenarios in our open-source platform to faster industries digital transformation. We are looking for partners to join our ecosystem together to deliver solutions to different industries together.
Please contact coope[email protected] for more partnership ecosystem information.