It was not long ago when we, the Silversparro Team, visited a large steel mill to study its processes that make it successful and profitable. To our surprise, for every process and group of workers there were a dozen pairs of eyes supervising and monitoring them closely. During the time when we expect every part on the factory floor: machinery, equipment and processes to be fully automated, manual operations still dominate the manufacturing sector.
“So, what do you think about our Supervision Army?”,
said the manager.
“We have spent a handsome sum on employing the personnel well-versed and well-trained in their respective fields,”
We learned that every manufacturing facility has its own army that is deployed on the factory-floor. The only difference here is that the army sometimes attacks the wrong region that further adds to the danger prevalent on the floor. The army is not to be blamed here as humans tend to make mistakes and with a blink of an eye, a small but potent issue can be missed.
“Last month, we had a machinery breakdown exposure which was, of course, unpredictable and unnoticeable and it has affected the steel production. We have been suffering loss since then.”
Surprisingly, manufacturers are aware of this incompetence of humans to monitor and supervise everything but not of the solution.
“Have you tried looking for an alternative?”,
asked our team member.
“Yes, we have had many discussions and every time we ended up with one disappointing conclusion. ROBOTS!!!”,
answered the manager, jokingly.
The manager didn’t realise that they’re close to the solution as owing to artificial intelligence and its combination with machine-learning algorithms and edge computing all common shop-floor problems can be avoided and resolved.
“We have even considered installing sensors in our system that would send an alert on any abnormal activity.”
said the manager.
“But a given sensor is designed for a specific purpose and you’d need a sensor for every process and equipment for monitoring different purposes.”
replied our team member.
Sensors can be considered a good solution but one incurring an enormous cost . Every sensor is designed specifically to collect data for a specific purpose and equipping your production system with sensors will invite a hefty amount of money.
So, sensors will not be sufficient for the monitoring and supervision of processes and people. In order to solve the problem of supervision the aforementioned solution needs to work in harmony with the combination of artificial intelligence, machine learning and edge computing. This combination has given rise to the evolution of Video Analytics which is driving workplace productivity in manufacturing.
Video analytics can be defined as the process of detecting movement or motion with a fixed background. It is capable of generating automatic alerts at the time of any abnormal activity in the frame. These mathematical algorithms monitor and analyse both real-time and post events. Moreover, the video analytics market is predicted to grow globally at an exponential cagr of 22.67% during 2018-2026.
“It’s good to see your plant with all kinds of machines. Also, the manufacturing industry has entered the fourth revolution now, are you ready to adopt this change too?”
asked our team member
“We‘re not sure. We are already at a loss and can’t afford to risk our production anymore.”
answered the manager.
During the third industrial revolution, computers and automations were introduced and incorporated in the manufacturing processes. Industry 4.0 has enhanced and advanced the preceding revolution as machines and equipment, connected over a network, are communicating and interacting with each other while taking the relevant decisions.
Manufacturers still seem skeptical and are struggling to adopt Industry 4.0 as it not only changes the way machines work but also the way humans work. Like the manager, there are many others who can’t identify the right opportunities to embrace the future. As Forbes quoted: “Since connected machines collect a tremendous volume of data that can inform maintenance, performance and other issues, as well as analyze that data to identify patterns and insights that would be impossible for a human to do in a reasonable timeframe, Industry 4.0 offers the opportunity for manufacturers to optimize their operations quickly and efficiently by knowing what needs attention.” No business wants to be a laggard, it’s better to be an early adopter and reap the benefits rather than being a laggard while watching others win the race.
However, the majority of manufacturers are curious to see the wonders Industry 4.0 will bring to their business. According to a report from Capgemini Research Institute, over half of European manufacturers (51%) are implementing at least one AI use cases in the sector where Germany is the frontrunner with 69% of AI adoption. AI is believed to transform the industries in terms of reducing the operating cost, enhancing the quality and improving productivity.
“Many manufacturers have attached sensors to their devices that have transformed the traditional manufacturing to a more powerful and interconnected system”
explained our team member.
“And with IoT, video analytics has gained strong grounds too.”
“So, as per IoT all the machines should be connected over a network and that will resolve most of our problems then how can video analytics help us?”
asked the manager.
If we compare Internet of Things (IoT) applications and video analytics applications then the latter provides more value. According to a Report ‘Video meets the Internet of Things’ by Mckinsey & Co., IoT applications usually offer more value when they incorporate video analytics, since the technology allows them to consider a wider range of inputs and make more sophisticated decisions.
AI based video analytics is at its tipping point as according to a report by Technavio, video-analytics is showing signs of growing more and more prevalent in the world, predicting an 11% growth in the global video surveillance market between the years 2019 and 2021.
Find out what causes the sudden surge of interest in harnessing video analytics in your businesses here.
“Does that mean video analytics can solve supervision problems at our manufacturing unit?”
inquired the manager.
“Yes of course, our new tool Sparrosense AI Supervisor is your friend on the factory-floor”
replied our team member.
As the name suggests, a supervisor who looks over all the factory-floor activities using the power of artificial intelligence. Silversparro has launched Sparrosense AI Supervisor, a tool developed using deep-learning algorithms, is your friend on the factory-floor. It works with the existing DVR systems and CCTV cameras paired with video analytics software that act as IoT sensors and helps manufacturing units to optimise operations, protect people and proliferate profits.
It automates all kinds of manual supervisory work on the factory floor using the power of video analytics. Manufacturing units that are home to numerous moving parts like people, machinery, raw materials and equipment can be benefitted by deploying it on factory-floor. In order to produce the desired product all these parts should work together seamlessly. Apparently, processes that include these moving parts and components are prone to have glitches and resolving them on the basis of manual observation can sometimes escalate the problem.
AI Supervisor acts as a friend in terms of reducing operational cost, improving the safety of workers, enhancing the quality and augmenting the revenue ultimately adding the word ‘smart’ before your factory. Hence, making it a smart factory.
It captures processes using CCTVs and then smart AI algorithms identify processes using visual cues. Captured footage is then processed to detect both machine movement and worker actions. AI algorithms help predict delays and if it detects any then it raises alerts in real time while recommending actions required to manage the delay.
“Sparrosense AI Supervisor not only monitors human actions but also every process in order to streamline the production”
stated our team member.
“What kind of processes does Sparrosense monitor?”
another question by the manager.
Is it just the manager who wants to know the processes that Sparrosense can monitor? Of course not, you too wish to know how it can help you and your business.
Intelligent AI algorithms on CCTV systems makes the environment safer, smarter and stronger. The processes for which these algorithms alert the relevant stakeholders of any abnormal activities are as follows.
“Now, we’ll have to buy new CCTV and DVR systems. This will result in more cost”
said the manager.
“What if we tell you that your existing infrastructure will be put into service”
our team member responded.
Almost every manufacturing facility now-a-days is installed with CCTV cameras. According to the Global CCTV Market Forecast 2022 Report, analysts have identified the CCTV market to grow globally at a CAGR of around 11% during 2018-2022. As the statistics show a tremendous increase in the CCTV market due to security reasons, the same CCTV cameras can be used to enhance productivity, improve the safety of workers, reduce operational cost and increase the security of machines and equipment by using the power of video analytics. Evidently, the demand for IP-based cameras is increasing, and is likely to boost the demand for video analytics appliances.
CCTV cameras are installed in a manufacturing unit for site surveillance and monitor performance of personnel, machinery and equipment. Since, each camera has a defined area of recording and after the intelligence is added to that particular camera, it knows what happens in that area ‘normally’ and can decide to raise alert if detects something suspicious.
For instance, a camera captures the workers in a particular area wearing a hard hat (safety helmet), a worker enters the scene without the hard hat then the video analytics software makes the camera raise alert to the management. Eventually, the management will take the necessary actions.
“All of it sounds interesting to me but what about the workers, when half of their work will be managed by the Sparrosense?”
exclaimed the manager.
“It is not the case. The personnel will always be the real decision makers and real action takers”
answered our team member.
In the aforementioned example, management is doing the real work, it is the real decision maker and actions taker. AI Supervisor will not replace humans but change and augment their jobs. It has reduced the burden of ogling and monitoring x number of screens.
Who were once skeptical of adopting the new transformative technology, AI, are now waiting in the queue of becoming the frontrunners. Earlier manufacturers feared the changes AI technologies would bring to their processes as it promises to change everything from business models to operational models to how the work is done. However, AI is changing everything for the better.
The scenario seems to change in the near future. Until now, manufacturers used to expend a whopping amount on training a personnel who would sit in a room with a dozen screens before him to monitor and supervise the workers’ and machines’ activities. But now, the huge amount of data generated through CCTV is processed using smart AI algorithms hence, saving the extra money, efforts and time most importantly, improving the quality and accuracy.
PTCIL is a leading steel manufacturer that was facing issues in Mold Preparation and Furnace Melting which resulted in 50% less production than the capacity. Smallest of delay and lack of coordination was leading to immense loss of productivity. Sparrosense was able to accurately measure and predict delays just by analysing cctv feed and automatically alerting workers. This unlocked the bottlenecks driving revenues significantly.
This manufacturing plant had issues with three of the setup processes Launder Insertion, Top Lance Positioning and HM Positioning which have to function in parallel. Delay in any of the process resulted in less time available for more heat, hence leading to overall delay in production. Sparrosense identified the status of all three processes and the data collected was displayed to the management in real time. Moreover, it sent alerts to workers, supervisors and managers at the time of delay in any step.