Computer Vision in buildings

Computer Vision (CV) is a subfield of artificial intelligence that aims to extract information from visual data. Photos, videos or even multi-dimensional data, such as from time of flight sensors, can serve as the source for the evaluation. CV is based on the human ability to capture, process and analyze images. In some areas current machine learning algorithms are already so good that they have surpassed human performance. We develop sensors for the detection of people, objects and vehicles in buildings with the goal of optimizing the flow of people in buildings. Here you will find an overview of tasks, application areas, technology and possibilities of the sensors.

Optimization of the elevator control

The CV-API can provide information about the occupancy of the elevator car, the number of people waiting on each floor and special objects (beds, robots, goods, cars) that are in the elevator car or waiting to be transported on a floor. If this data is available to the elevator controller, it can optimize the trips using suitable algorithms and thus shorten waiting times and increase transport performance.

For example if an elevator cabin is occupied by a larger object (e.g. bed, goods) or many people, then no further people can enter the elevator. With the data provided, the elevator controller can decide to drive directly to the destination floors and not process the calls from waiting persons until the cabin is empty again. This avoids unnecessary stops and increases the transport performance.

Bettenfahrt mit Computer Vision erkennen
Hand trapped in elevator door
Increase safety

Elevators are among the safest means of transportation. Nevertheless, accidents sometimes occur, mostly due to incorrect behavior of users or service personnel. The CV-API can be used to provide information that can help increase safety.

People or objects moving quickly towards the closing door can be signaled to the elevator controller, thus preventing entrapment. In glass elevators, children standing with their hands on the elevator door can be detected, thus preventing their hands from becoming trapped when the doors are opened.

Computer vision for emergency in elevatot
Functions during emergency call and in case of fire

If an emergency call is received from an elevator at an emergency call center, a voice connection is automatically established. If no one then answers, the service center does not know whether the person has left the car or lost consciousness and an emergency situation exists. The CV sensors can detect whether the cabin is empty, whether there is at least one person in the cabin and whether they are lying motionless on the floor. This is valuable information for rescue. In the event of a fire, this information can also help firefighters on the rescue floor or the fire alarm panel to quickly find and free trapped people. In addition, 'Smoke and Fire Detection' can provide additional information about the spread of the fire.

Computer-vision-robtos-in-elevator
Sensor values for special trips

Elevators are often also used for transporting goods or vehicles. Certain regulations must be observed for this, e.g. that these objects may not be transported at the same time as people. The data from the CV sensors inform the elevator controller whether there is a robot, a bed in a hospital or a vehicle in the car so that a direct trip to the destination floor can be carried out automatically.

Computer vision - persons in wheelchairs
Door time control

The doors of elevators very often close after a short time to allow high transport capacities. However, elderly persons, wheelchair users or parents with strollers often need a longer time to enter or leave the elevator. A door that closes too early can then cause an accident. The CV sensors can detect wheelchair users, strollers and elderly people moving very slowly and keep the doors open longer.

Conclusion

Computer Vision also enables elevators to increase safety, transport performance and comfort in a variety of ways. Depending on the application, different sensors must be used. When planning the project, in addition to the customer's requirements, the lighting conditions, viewing angles, ambient conditions and interfaces to the building's technical trades must also be considered.

Persons with suitcase in elevator - airportComputer vision for hospitals

Product Groups
Products Tags