Secondary monitoring is considered to be income generation monitoring that provides inferences from various activities taking place in stores
By Atul Rai
Artificial intelligence-based video analytics has gained ground to the point of becoming buzzwords of the 21st century. From navigation and safety, to facial recognition and even vehicle analysis, AI-based technologies have revolutionized our lives and continue to amaze us. According to a report, more than a billion cameras were sold last year. In addition, Indians sold 30 million CCTV cameras in 2019-2020 with a CAGR of 20%. Research indicates that we have more cameras than humans on earth. The automation of CCTV surveillance is therefore certainly beneficial in several ways.
We install cameras widely for activity monitoring, vehicle monitoring, people monitoring and object monitoring. And this surveillance is carried out by humans themselves. The challenge with human monitoring is that we cannot focus on a particular screen for more than 20 or 30 seconds due to various psychological facts. So the problem is, we depend on human capacities for surveillance and humans are not even efficient at it. So this is where automation is needed. Camera surveillance is divided into two parts –
1. Loss prevention or direct monitoring
2. Secondary or income-generating surveillance
Direct monitoring that we see in manufacturing, infrastructure, retail or government organizations where activities like accident monitoring, violence monitoring, cleaning monitoring, people detection are monitored. In addition, facial recognition where we do identification, whether for presence or to detect blacklisted / whitelisted people, is also part of direct monitoring. Its purpose is to prevent loss and is an economical method used for monitoring. For example, if an accident occurs with monetary loss or any violence will result in in-store sabotage. This is avoided by sending alerts in real time so that the necessary actions are taken to prevent any loss.
As for secondary monitoring, it is considered as income generation monitoring that provides inferences from various activities taking place in stores. Stats like how many people visited the store, particular clothes they were interested in, time of day the store was hiked, gender of attendance, etc. are collected for better decision making. Although video analytics started with an economic surveillance method, it also works for the income generation part. In brick and mortar stores, cameras are installed and security guards are also appointed to prevent any kind of intrusion or illegal activity. Thus, an investment is made in the infrastructure of the cameras and another in the human force to monitor it. Since the world’s population is huge, you can’t put humans behind every surveillance to avoid security holes. This method of monitoring should be automated. Let us take the example of the manufacturing sector where the checks of safety vests, the time of loading / unloading, the entry / exit of a vehicle, theft surveillance or we can say that all the operations of the l he manufacturing ecosystem and security issues are addressed using AI-powered video analytics.
Take another example from the retail industry. Currently, the competition of retail stores is done with e-commerce. Ecommerce has the luxury of all data with them. One of the main reasons e-commerce websites are winning the retail race is their ability to track daily customer inflows, stay time, and purchasing choices through an automated system. They know exactly what to get their customers the next time they open their virtual store.
Hence, some key customer analyzes which are crucial in making important decisions for marketing, operations, inventory control, etc. are not available with offline stores. Offline stores have blind sensors that are unable to give information such as when the store is getting more traffic if a store has invested in a planogram for occupancy analysis, what outlet a store is getting, customers actually visit the strategic aisle or not, demographics of customers visiting a store, etc. Not all of this information is about offline stores and they are blindly investing money in marketing. With video analytics, retail operators get this data, which makes their marketing campaigns more mature and generates more revenue.
The camera is the interpretable sensor that most closely resembles a human brain. For example, the sensor can detect a fire and to authenticate the alert generated by the sensor, a person has to visit that particular site, but a camera issues an alert, all that is required is to see the image, and the necessary actions are taken regardless of your location. The feature that makes CCTV surveillance very powerful and economical is that it uses the existing infrastructure and makes the cameras smarter by using the top layer of artificial intelligence which is video analysis. Thus, the adoption of this technology is easy because it does not incur any additional cost.
Additionally, whatever results generated that can be identified by humans are not possible with any other sensor. For example, any sensor that counts the number of objects loaded / unloaded on a truck, how will this be verified unless humans do not verify it.
This is the advantage of the camera. Moreover, a camera can do several things, the same camera identifies violence and also counts objects. The camera is ubiquitous, making it the best tool for automating various solutions. In summary, we can say that artificial intelligence-
Optimized video analysis has come to a stage that can revolutionize anything and everything and bring industries to exceptionally high ROI.
(The author is co-founder and CEO of Staqu. The opinions expressed are personal.)
follow us on Twitter, Instagram, LinkedIn, Facebook
Financial Express is now on Telegram. Click here to join our channel and stay up to date with the latest news and updates from Biz.