Safety hazards in every workplace come in many different forms: sharp edges, falling objects, flying sparks, chemicals, noise, and other potentially dangerous situations. Safety regulators such as Occupational Safety and Health Administration (OSHA) and European Commission often require that businesses protect their employees and customers from hazards that can cause injury by providing them personal protective equipment (PPE) and ensuring their use. With Amazon Rekognition PPE detection, customers can analyze images from their on-premises cameras across all locations to automatically detect if persons in the images are wearing the required Personal Protective Equipment (PPE) such as face covers, hand covers, and head covers. SAP customers use SAP Environment health and safety module to record these detections manually as safety observations.
Our solution provides an integration framework between Amazon Rekognition and SAP Environment, Health and Safety (EHS) and adopt the open-source Events-to-Business-Actions Framework, which will automate the process of creating safety observations.
In the upcoming sections of the blog post, we will describe the architecture, systems involved, PoC scenario and a quick demo, followed by the instructions to deploy the solution from a GitHub repository
The outcome described in this article was facilitated by an engagement between SAP and Amazon Web Services, which resulted in the creation and validation of a joint reference architecture use cases and PoC. Teams from PVN PavanKumar (SAP Labs India) and Soulath Khan (Amazon Web Services) have participated in this engagement.
Businesses today need to record and manage incidents, safety observations, and near-misses to ensure a safe workplace. Safety regulators like OSHA and the European Commission require businesses to provide personal protective equipment (PPE) to keep their employees and customers safe. But most of these safety observations are recorded manually and often happen after a safety hazard or an incident has occurred.
To tackle the tedious task of manually recording these safety observations, we have developed a solution that uses the Amazon Rekognition service to capture and detect if the PPE is missing, and in such a scenario happening in real-time, it generates an event that now needs attention and is sent to the Events-to-Business-Action Framework, which maps it to an associated business action, which in this case is to create a safety observation in the SAP S/4HANA.
With Amazon Rekognition PPE detection, customers can analyze images from their on-premises cameras across all locations to automatically detect if persons in the images are wearing the required PPE such as face covers, hand covers, and head covers. Using these PPE detection results, customers can trigger timely alarms or notifications to remind people to wear PPE before or during their presence in a hazardous area to help improve or maintain everyone’s safety. They can also aggregate the PPE detection results and analyze them by time and place to identify how safety warnings or training practices can be improved or generate reports for use during regulatory audits.
SAP Environment, Health, and Safety Management is a critical part of SAP Sustainability suite of solutions, including SAP S/4 HANA, SAP product footprint, and the SAP Sustainability Control Tower.
For more information about the SAP EHS solution, refer to the SAP Environment, Health, and Safety Management .
The business value of this integration is reducing workplace incidents and improving worker safety using artificial intelligence and SAP sustainability solutions.
High-level architecture (with SAP S/4HANA on AWS)
The following steps depicts the information flow across systems:
(1) Equipment (Camera): CCTV cameras at the factory capture images of people and location and pushes them to an Amazon S3 bucket.
(2) Amazon S3 holds the images for model inference. A lambda function can be scheduled or triggered to execute the inference.
(3) The lambda function calls Amazon Rekognition PPE detection model to detect any visual inspection issues with the product.
(4) Amazon Rekognition – PPE detection evaluated the image for any missing PPE equipment like Hand gloves, Face covers, Head gears and safety visors. if an anomaly is detected, The Lambda function accesses AWS secrets Manager for credential information for SAP BTP
(5) The lambda function calls the SAP Integration Suite, Advanced event mesh and passes the event payload. Event-to-Business-Action framework (extension app) processor module’s endpoint is subscribed to SAP Advanced Event Mesh, hence receives this event.
(6) Event-to-Business-Action framework(extension app) processor module leverages the Business Rules capability of SAP Build Process Automation to derive business action (for example, In this scenario,EHS Incident Report Safety Observation creation in SAP S/4HANA system) based on certain characteristics of incoming event.
(7), (8), (9) (10) and (11) Event-to-Business-Action framework (extension app) processor module triggers the defined action in the SAP S/4HANA system by using the SAP Destination Service and SAP Private Link Service.
To get the entire scenario working, there are prerequisites and initial configurations required in SAP Business Technology Platform, AWS, and SAP S/4HANA.
Please refer to the published GitHub repository and Discovery Center mission, which contains code samples and instructions for implementing this scenario.
We have created this reference architecture with accompanying sample code which can be configured and extended as per your requirements. However, it is up to the end user to utilize the code, make modifications and keep it maintained for productive use.
The business value of this integration is the following:
On-premises cameras constantly capture the workers on-site. The images are processes by Amazon Rekognition, which detects if any person has a missing PPE (which in our case is a missing hand glove) and sends this event to SAP Integration Suite, Advanced Event Mesh. The event is then consumed by Events-to-Business-Actions Framework which determines the associated business actions and Creates a Safety Observation for missing hand-glove in SAP S/4HANA.
We hope this blog has given you a brief idea of how to integrate events generated from industrial equipment to trigger different business actions, workflows, and processes in SAP backend systems.
On the SAP side, Many thanks to Uma Anbazhagan from SAP for driving this topic by collaborating with all the stakeholders involved in the engagement, Praveen Kumar Padegal and Ajit Kumar Panda for supporting the development of this solution, Madankumar Pichamuthu for enabling collaboration, and finally Anirban Majumdar, PVN Pavan Kumar, Sivakumar N for leadership support for this project.
On the Amazon AWS side, Many thanks to Soulat Khan, Ganesh Suryanarayanan, Sunny Patwari, Renga Sridharan, Aiyappa Machanda, Abhik Ray, Suresh Pulivarthi, Krishnakumar Ramadoss and the SAP APN Partner Innovation team for their support in this project.
To learn more about SAP BTP, see the learning journey on SAP Learning called Discover SAP Business Technology Platform, a great introduction to BTP and the Intelligent Enterprise strategy, to see what it’s all about for free.
As part of the global partnership between SAP and AWS, we have jointly developed an openSAP course titled “Build Resilient Applications on SAP BTP with Amazon Web Services
[Registration link] https://open.sap.com/courses/aws1
By joining this free openSAP course, you will delve into multiple Joint Reference Architecture patterns, gain in-depth knowledge of the architectural patterns, and have the opportunity to enhance your skills through hands-on exercises.
To learn more about Amazon AWS, visit AWS training
For more information about this topic or to ask a question, please contact us at [email protected]