/dq/media/agency_attachments/UPxQAOdkwhCk8EYzqyvs.png)
Follow Us/dq/media/post_banners/wp-content/uploads/2020/05/3C14974E-6156-4980-8730-6CBC4303CFED.png)
The food industry has grown tremendously over the past few years. The increasing demand for cooked and processed food, lead to an increase in the number of restaurants and food processing companies. With that, there have been rising concerns about the maintenance of hygiene during the cooking and packaging process of the food. Bad hygienic practices can lead to serious health issues like food poisoning, for the consumer, which can, in turn, lead to lawsuits on the company/restaurant.
Traditionally, these practices have been monitored manually by food inspectors and supervisors. However, due to human error and bias, it has always been preferable to have an automated system to ensure hygiene and safety compliance.
The rise of AI-driven Video Analytics
The concept of Artificial intelligence has been there since the 1950s, however, the availability of data and computational capabilities were the major limiting factors at that time. Fortunately, in more recent years, we have been able to overcome both these obstacles and now it is feasible to run AI-based algorithms to run Video Analytics in near real-time.
The Algorithms
Artificial Intelligence-based algorithms have flooded the technology space. Every year a faster and more accurate algorithm is discovered, which makes it even more feasible to run real-time video surveillance for safety and compliance in the food industry.
Due to government guidelines, every kitchen, and food-manufacturing plant is already equipped with CCTV cameras. Additionally, Computer Vision algorithms have become efficient enough to be run on real-time feed efficiently at minimal hardware cost. With a fast internet connection, the video feed can be securely streamed to a remote server for processing for heavier algorithms. Hence, most of the supporting infrastructure is already in place.
Video Surveillance in the Restaurants and Cloud Kitchens
Autonomous FaceMask, Hairnet, Apron, and Gloves Detection
As per regulations for proper sanitary practices, every employee who is directly in contact with food is required to wear appropriate protective clothing to avoid contamination of any kind.
Each of these is required to be worn at all times while near or around food. However, it is not possible for human beings to monitor the same during all working hours. Hence, this problem can be solved by state-of-the-art Object Detection models, which detect the presence of these items at all times. In their absence, an alarm can be raised.
/dq/media/post_attachments/wp-content/uploads/2020/05/5ED46229-24DC-4FE9-87AD-8F5D6A7F6CCD-300x169.png "ED DC FE AD F D A F CCD")
Autonomous calculation of frequency and duration of handwash
Clean hands are a fundamental part of hygiene when it comes to cooking/preparing food. To ensure this, it is necessary to ensure frequent handwash for a minimum duration. At a human level, it is impossible to monitor and calculate the duration and frequency of handwash for each employee.
However, with state of the art Activity Recognition models along with Facial Recognition, the duration and frequency of handwash can easily be monitored for each employee. The same system can be used to alert the employee if the minimum duration of handwash has not been completed through real-time audio or visual alerts.
Video Surveillance in the Food Processing Industry
Productivity Monitoring
Since the food processing industry employs a large workforce, supervising it manually is difficult. Hence, activity recognition models can be used to detect the duration of activity performed by the employees to keep an account of productivity.
Additionally, another layer of Business Intelligence can be added on top of the existing activity model to determine which production line(s) are creating a bottleneck, so that they can be addressed accordingly.
Worker Safety and Hygiene Monitoring
Similar to the restaurants and cloud kitchens, maintenance of hygiene is important. Hence, for the employees coming in direct contact with food, gloves, aprons, and hairnet monitoring is important. In addition to that, worker safety is also a priority, and hence, for employees dealing with heavy machinery in the production line, hard-hat and protective equipment detection is also important.
Conclusion
Only a couple of applications have been discussed so far. However, the possibilities are endless. Smart Video Surveillance can be extended to detect the frequency of mopping and cleaning, detection of spillage or to monitor the cleanliness of a table surface.
It is clear that Smart Video Surveillance is the way forward. It can reduce the work of supervisors significantly as well as save the consumer from being served with bad quality or unhygienic food.
- Adit Chhabra, Co-Founder & CEO of Wobot.ai