Edge AI Vision For Factory Automation

Edge AI vision for factories could be a game changer in many domains and give that edge in improving not only their overall process efficiency but also improvement in Safety and Environment control.

Factories worldwide are now more focussed on improving efficiency in all aspects of operations thus improving their dwindling profit margins and at the same time being compliant with Sustainability requirements.

All available opportunities to meet the above objectives and remain cost competitive are being explored. AI/ Digital Solutions that is believed to bring in the next transformational impact is being widely considered emphasizes Rohit Kochar, Bert Labs.

Industries across domains are hoping that it shall change how traditionally factories have worked so far and bring in significant impact on costs as well as meet sustainability objectives.

IoT platform, Communication networks, Enterprise software, Distributed Computing and Artificial intelligence technologies are seeing an increased demand in Industries. The objective is to bring in increased efficiency in processes that are already optimized using earlier available technologies.

Solutions using existing available data or using IoT devices for data capture to feed AI models that shall then work their magic are now gaining popularity and have started to find use in many applications. Factory data usually consists of sensors data of the process like temperature, pressure, flow, valve positions etc.

These are numerical data. Visual data are captured in few processes like cameras installed for burner flame monitoring in cement kilns. But using visual data for AI computation to improve the efficiency is rarely heard of. It could be due to limitations in visual data capture and the computational power required to process the visual data.

Edge computing device equipped with image capture technology is gaining traction in factories to support and improve the existing Automation/AI systems. IoT Edge devices can latch on to Wireless Sensor Networks or connect by LAN/Ethernet cable for two-way communications and data transfer.

The data is sent to the remote server like cloud or on premise server where data processing or AI computation is done. Most of edge devices can infer the AI model’s output. But edge devices that can run AI computation in the device itself will change the way factories utilize AI/Automation systems. Edge Computing devices can be equipped with both digital camera and thermal camera.

Based on application, the camera specifications are determined. These devices can host powerful GPU enabled chips, that can process complex computer vison algorithms.

The advantages of Edge vision are manifold. Local data processing enhances data privacy and security, as sensitive information does not need to be transmitted to the cloud. Factories can easily scale their operations and adapt to new production requirements with the flexibility provided by Edge AI. By automating production processes, Edge AI improves operational efficiency and accuracy.

Edge AI processes data locally, reducing latency of cloud compute and enabling quick, real-time decision-making on the factory floor. Defect detection can be done real time, increasing yield and reducing waste.

In factories where temperature profiles in the process are very critical like Kilns in Cement/ Tile manufacturing, or ED Oven in Automobile Paint shop, the capture of thermal images and processing the images using computer vision algorithms can provide useful insights on the process conditions and equipment health.

It shall help Plant operators to take decisions that can prevent breakdowns and control the process in an efficient manner. Similar advantages are there for digital image captureand processing like Factory Gate monitoring for security and safety compliance.

Manufacturing Plants have isolated and remote areas which are not regularly under operator surveillance, Visonpowered edge computing device can play a vital role in such areas and notify if an abnormality like a gas leak, chemical leak, dust emission is happening at its nascent stage itself thus preventing any major environment event.

CCTV cameras can detect visual abnormality, but manual monitoring of CCTV feed needs to be there. In Buildings where temperature control is important, number of people inside the room affects the heat load.

Occupancy can be determined by these edge devices and processed output can be taken as input for the HVAC control thus saving energy, maintaining thermal comfort as well as meeting the process requirements.

Visual quality checks for many finished products now being done by quality inspectors. These can be automated using the Edge Vision device. Major advantage is that the processing can done at the device itself without sending the data to a remote server. This reduces the latency and reaction time.

Vision enabled IoT device can form part of a larger IoT device network that is capturing other data from the factory.

The combined data can be transmitted by IoT powered wireless network to a gateway that routes the data to remote server hosting big data lake software platform and multitenant distributed computing.

AI computation happens here, and powerful insights or optimization commands are generated. This will bring in much greater impact on the plant efficiency.

In a factory premises, vision enabled edge computing devices can have many other applications as well. In Product dispatch section, the incoming trucks and outgoing trucks monitoring can be automated.

Personnel entry in critical areas where only authorised personnel are allowed can be monitored and controlled. Safety and count of people working in confined spaces like reactor, vessels, columns can be monitored.

Hence vision enabled Edge AI can help streamline operations with high precision quality control, enhancing process line efficiency and accuracy for robotic guidance, monitoring inventory real time optimizing stock levels and reducing waste, predictive maintenance of equipment preventing unexpected breakdowns and reducing downtime and most importantly, monitor the factoryenvironment for safety hazards, alerting workers and management to potential dangers and ensuring compliance with safety regulations.

Opportunities are many for the utilization of vision-powered IoT devices. Important thing that factories should realize this at the earliest and adopt the technology as part of their innovation strategy.

Thus, Edge AI vision for factories could be a game changer in many domains and give that edge in improving not only their overall process efficiency but also improvement in Safety and Environment control. According to Rohit Kochar, Founder, Bert Labs, factories should embrace technological innovation like Edge AI vision and adopt vision enabled edge computing device to strengthen their AI/IoT initiatives and bring all round improvement in their operations.