CCTV Intelligence in Retail post COVID-19
Impact of COVID-19 on Retail
The COVID-19 pandemic has created unprecedented health and economic crisis across the globe. Strict limitations, quarantines, and restrictions have resulted in businesses shutting down in the past few months. Some industries like retail have suffered more than others. However, with the lockdown easing, store closures are coming to an end, and retailers are starting their preparation for the post-COVID-19 scenario in which stores will reopen. Social distancing rules are likely to be in place for more time and this is likely to mean only a specified number of people will be allowed into the store. Brick and mortar stores will undoubtedly be struggling to regain footing in an evolving omnichannel marketplace where retailers are balancing e-commerce, in-store, and click-and-pickup shopping. Retail businesses must start accommodating the needs of an increasingly digital world in which the in-store experience has changed forever.
CCTV advancement and ineffective use
Many shopping centers already have video surveillance cameras that are commonly used to track the security of their business operations. Monitoring each CCTV feed manually isn’t practical for security and surveillance. Monitoring multiple CCTV feeds at the same time is both inefficient and ineffective, as well as labor-intensive. Because of these shortcomings, CCTVs are reduced to a mere tool for post-violation analysis. With the advancements in technology, gadgets are getting more intelligent in capturing, analyzing, and disseminating the information, and cameras are no exception. With a large number of cameras already connected to the Wi-Fi or internal networks, we can efficiently work along with other connected devices for enabling intelligence even if there is no built-in intelligence in cameras. Artificial Intelligence can be effectively used to radically increase the efficacy of the video surveillance systems with the help of real-time alerts to avoid any security breach or potentially threatening situations in these stores.
How can AI solve the problem?
With the lockdown restrictions being gradually relaxed, the retail stores need to reassure their customers and employees that it is safe to work or shop. It will go a long way in attracting consumers and retain employees. As the global and local retail companies are grappling with post-COVID safety measures that necessitate a clean, safe environment where social distancing is maintained, artificial intelligence can restore normalcy with resilience. The idea is to brace up retail stores, malls with AI-enabled surveillance systems. Computer vision and video analytics technologies are very useful to facilitate safer and more efficient shopping experiences for both buyers and employees. We understand that retail stores would be having no interest in deploying and paying for additional infrastructure when most stores already have several cameras installed. We make this vision work with the already existing security installation systems of the retail stores. In this blog, we propose to analyze CCTV feed by implementing a layer of artificial intelligence that makes the system capable of solving the below-mentioned use-cases more efficiently than the human eye does.
1) Health Protection in Retail Stores
Considering how fast the COVID-19 outbreak is spreading and the plethora of government regulations in this regard, it is essential to have strict measures to prevent the spread of diseases and gain the confidence of customers. Even though contactless thermal scanners are useful, they rely entirely on human control. The usage of AI-led thermal scanners installed within the CCTV cameras can be considered for a contactless customer screening. The technology uses a combination of face recognition, thermal camera processing, and artificial intelligence to detect the body temperature of multiple people at the same time and identify the outliers. We can also use AI to ensure social distancing is maintained. Individual people can be identified from the CCTV feeds with the help of the person detection technique and their proximity to other individuals can be calculated using 3D distance calculation techniques. If certain individuals are found to violate social distancing norms, alerts can be sent to the store manager along with their identification. We can even detect the customers who are not wearing the mask and provide them with the same. As cleaning and sanitizing has become very important to check the growth and spread of the virus, we can track the housekeeping staff if they are cleaning at regular intervals using action detection and person tracking techniques. With all the things getting automated, we can continuously inform the store manager and staff in case of violations in real-time.
2) Crowd Control and Queue Management
Since the onset of the COVID-19 pandemic, managing crowds has become problematic. But with the advancement in AI, we can automate the process of crowd control without any manual effort. We can use AI on the CCTV feed to get the footfall/person count in different areas of the retail shop and avoid overcrowding by alerting the store manager using pre-defined safe occupancy thresholds. We can also use the CCTV feed at the entrance and exit to set up a “one in, one out” policy to control the crowd by using a person detector. For example, we can notify waiting shoppers when they can enter a store via a “traffic light” device. Active queue management is critical for organizations in the present time. We can ensure a safer in-store experience with shorter wait times in the checkout line. We can use an AI-enabled person tracker to analyze the length of the queue at each billing counter and the time taken to check out purchased goods from the billing area. In case of a long queue in the billing area, considering the store’s wait time policy, we can alert the store management to deploy additional staff immediately to alleviate that congestion. By continuously tracking the billing areas, we can deploy staff at the busiest spots on time, reducing frustration for a time-conscious shopper while enabling the store to use available staff resources effectively.
3) Video Analytics and Heatmap
As the customers in the post-COVID world are moving towards online shopping, it’s essential to provide better in-store customer experience to attract them back to the retail stores. To do this, we need to collect more information about the customers and provide a better shopping experience. Video Analytics, which essentially includes videos from CCTV feed being read and analyzed with the help of deep neural networks (AI), can enable the retailers to have a profound impact on their sales curve. We can use the deep neural networks for identifying the age range and gender of a person that can further help us in better understanding our target customers. We can also track the customers’ paths, object interactions, and dwell durations using deep learning-enabled person tracking and object detection. Retailers can have a heatmap of the shop based on the time shoppers spends in specific areas, identifying the hot and dead zones of the store. We can use them to optimize customer traffic flows through shops and design more effective floor plans to offer a streamlined and satisfactory customer experience.
Conclusion
The difficulties involved in returning to the new normal is going to test the resilience and adaptability of the entire retail sector. The main priority is not just going to be disinfection of the stores but also the security so that customers can shop in a safe environment and their confidence can be won. Although, there is no lack of security cameras in stores, using them proactively and enabling them with AI is going to be necessary. Video Analytics is the next step in the extraordinary journey the retail industry is experiencing that will be enabled by Artificial Intelligence. For retail businesses, there is a fantastic opportunity for using video data along with operational data that will enable them to deliver transformative business outcomes and to answer the most difficult strategic questions
