In an industrial environment, technologies have other aspects. The cameras allow you to scan the temperature of the room, individual devices and material and set the parameters of the environment as needed. Machine vision is an invaluable part of safety solutions that go beyond conventional machine safety. Unlike light or radar barriers, they prevent people from colliding with machines in any environment, which is invaluable especially where autonomous transport robots come into play.
Occupancy of space
The common theme of energy savings, health protection and safety is the occupancy of buildings. The topic gradually grew with a modern approach to office work, it became urgent with this year’s pandemic. For the correct setting of the building and ensuring health, it is no longer enough to indicate how many people passed through the turnstile out or in. Occupancy data is important for each floor, sector, work cell, including constant monitoring of whether people maintain the specified spacings. As always, there are a number of perspectives on this:
- Building management needs to know the parameters for efficient heating, air conditioning and air exchange.
- Security needs to know the number of people in case of an evacuation.
- In order to protect health, it is necessary to at least draw attention to the excess number of people.
- The building operator or employer needs to protect and document that he is complying with the regulations for a specified number of people in one place.
It has been mentioned several times that the solution is machine vision. First of all, it is not possible to imagine the described technologies in the required level of reliability when using the image of only one camera. It is necessary to exclude blind spots, overlaps of people or objects and their clear distinction. This requires simultaneous image processing from several sources, their precise synchronization, and job processing at a predetermined time by type. There is a clear parallel with the systems developed for autonomous cars. The more weight the security task, the faster its output must be. Artificial intelligence is important not only for speed, but especially for accuracy. For each activity and image, there are dozens or hundreds of possible scenarios. Only then can we talk about an intelligent sensor.
Artificial intelligence plus machine vision
Artificial intelligence, supported by machine vision, has clear preconditions for itself. It becomes the basis for the management of next-generation buildings using information on the movement of people inside. In addition to deployments in the form of individual systems, companies are already working on integration that combines technical and security applications. The result will be a lower density of cameras, data stream, stored data and ultimately lower system cost and consumption.
The development is intensive. The main driver is the Asian market, where, in addition to traditional technology leaders, companies from India are also gaining ground. Many companies in the US and global players are focusing on the same use of artificial intelligence.
The advent of machine vision is also working in companies involved in the development of subsystems. At the same time, it is clear that machine vision technology and artificial intelligence subtasks are available to a wide range of developers, including modules for popular open source platforms.
