This content was originally posted in the "Insights" section of Azena's blog:
Noema’s Seat Occupancy Monitoring computer vision application can be used by transportation and safety authorities to monitor seat occupancy on public transportation and in other seated environments. Using human behavior tracking, the app applies computer vision to positively identify the given occupancy rate of seated areas and account for corner cases in which other systems fail.
By providing better real-time data for controllers in transportation and queuing environments, Noema’s Seat Occupancy Monitoring app can help increase the efficiency and accuracy of ticketing systems, passenger loading, customer service, safety and security checks, and analytics.
Transportation efficiency and capacity have a direct effect on transport costs. The data provided by the Noema app can be used to seat passengers more efficiently, maximize cabin utilization, reduce empty or low capacity backhauls, and better address the needs of handicapped customers.
The Problems with Current Solutions
One of the most widely used real-time solutions for seat occupancy monitoring is pressure sensors. While straightforward, pressure sensors come with several logistical and analytical drawbacks that are directly addressed by Noema’s computer vision solution.
Pressure sensors are incapable of determining whether a seat is occupied by a passenger or some other object, such as luggage or pets. While inexpensive, a sensor is required for each seat, and maintaining an accurate system requires testing and servicing each sensor periodically. This is costly both in terms of hardware and labor costs.
In a seated queuing environment, computer vision provides additional real-time queue data for ticketing systems, which themselves can’t accurately account for customers who take more than one ticket or vacate the queue.
Noema’s Seat Occupancy Monitoring app uses a single ceiling or wall-mounted camera to monitor the seats in an entire transportation cabin or in other seated environments, providing a live video output and easily digestible metadata. It has been trained to work in buses, trains, streetcars, and artificial seating environments. Additionally, the app has been trained to work with fisheye cameras, further reducing hardware costs by minimizing the number of cameras needed for a given environment.
Because the app utilizes human behavior tracking to understand the actions of passengers, Noema’s app is robust against common corner cases, such as luggage or other objects in a seat or passengers taking up more than one seat. Because of this, the app provides the best real-time occupancy data and statistical counting data for transportation environments.
Ease of Setup and Integration
Setting up the app is simple. Once the camera is mounted, the configuration is performed remotely. The user simply draws bounding boxes around the seats to be monitored, assigns labels, and begins monitoring seats. There are also user-friendly settings for adjusting the detection parameters and fine-tuning. If the configuration is going to be used in common, repeated environments (such as on a fleet of vehicles), configuration files can be exported to other cameras.
The app produces easily digestible metadata based on the Azena app store standard. The app translates metadata into the ONVIF XML schema, supports Message Broker and Data Trolley, and can be easily integrated into other 3rd party message brokers with the help of the Noema team.
Noema’s Seat Occupancy Monitoring application is the ticket to robust, real-time occupancy information for transportation and safety systems. With fisheye compatibility and remote configuration, our app drastically reduces hardware and personnel costs related to system maintenance, all while providing higher-quality, real-time data thanks to our human behavior monitoring expertise.
Noema is a computer vision products company that creates apps that perform object detection, video output modification, and human behavior tracking for a variety of industries. Noema is the owned subsidiary of DSR Corporation and was born out of our market-leading expertise around porting complex computer vision algorithms to edge devices.
Noema is actively seeking partners who are eager to test and deploy computer vision in a way that will revolutionize their industries. You can find our Seat Occupancy Monitoring application, along with other innovative apps from our portfolio on the Azena App Store. Download the app today and see why Noema was selected for first prize in the Azena App Challenge 2021.