How computer vision algorithms work
The most sophisticated computer vision algorithms are based on a kind of artificial intelligence known as a convolutional neural network. A CNN is a type of artificial neural network and is most often used to analyze visual images.
What computer vision algorithms bring to the table is scalability and the ability to memorize results. Instead of capturing and storing large amounts of video data, computer systems can, for example, store the fact that a person was present in a certain location at a certain time and was wandering from point A to point B .
Computer vision can automate the process, extract that metadata out of a video image, and then store the metadata without the image.
MORE FROM STATETECH: How is artificial intelligence used in state government?
Computer vision applications: Object recognition
Some applications are better suited to the eyes of a machine. âAI vision is ideal for applications where stable and reliable visual coverage of the scene of interest is possible and where there is a reasonable understanding of the range of types of objects or activities that will be present during use, âsays Mike Piacentino, senior technical director of vision systems for SRI International, a nonprofit scientific research institute.
This is why traffic management is a natural growth area for the deployment of intelligent vision systems.
GRIDSMART uses a single fisheye camera that provides a complete view of intersections, including the dangerous center where cars, trucks, motorcycles, scooters, bicycles and pedestrians intersect. âWe combine traditional computer vision and deep neural networks,â says Gabriel McFadden, senior regional sales manager for GRIDSMART. âGRIDSMART detects and tracks all moving objects in the scene and uses AI to determine exactly what they are. “
The platform provides real-time performance data such as time-stamped traffic volumes, turns and average speeds for intersection light analysis and synchronization and feeds the information back to traffic managers. GRIDSMART cameras are currently installed at 10,000 locations across the United States, including Tacoma, Washington, Seattle, San Francisco and Reno, Nevada.
GRIDSMART’s cameras use AI to track objects at intersections. Source: YouTube
John Rader, a senior traffic light electrician in Tacoma, says it typically takes four cameras to monitor an intersection with a light signal. A single GRIDSMART camera provides a 360 degree view, and some can tilt and zoom.
âIt worked really well for us and it’s easy to put together,â says Rader. âThere are a lot of features out there that we love. This includes some improvements that can now detect cyclists and pedestrians and give them more time to cross. Rader says the city is also using the system to count cars. At one intersection, planners were surprised to find 45,000 cars driving in a single day. âThe reporting is strong,â he says.
Computer vision applications: image processing
Hayden AI offers a smart camera for use in buses, police cars and other municipal vehicles. The on-board device can detect if a car is parked on a bus lane and can also mark and report parked cars at bus stops.
The system also stores historical data information to produce 3D maps. For example, it can detect when a car is parked in front of an invisible fire hydrant because the camera has noticed the hydrant on previous passes. The camera is a 21 teraflop device and captures data in 10 second evidence packets, which are transmitted to the cloud.
âWe are witnessing the digital transformation of the police,â says Chris Carson, CEO of Hayden AI. âEurope has done a better job with it. We cannot hire enough police officers. Many officers are concerned about COVID-19 or the danger of stopping a vehicle.
âOur technology has no bias; it just recognizes a car, âCarson says. Hayden AI has municipal clients in New York, California, and Washington State. Its cameras will also soon be found on city vehicles in Washington, DC, Carson says.
RELATED: How can cognitive engine technology contribute to public safety?
Computer vision applications: motion tracking
Intelligent vision systems also find their place in other specialized applications that can be used by state and local governments. Smart Vision Works is a company founded in 2012 with clients primarily in the food and agriculture industries.
Smart Vision Works collaborated with the State of Michigan on a pilot project to identify and remove invasive carp from its waterways. The native fish at risk is called the June sucker. The company trained its AI system with deep learning to identify the fish among six or seven other species.
âYou don’t know exactly what features it relies on,â says Beau Tippetts, vice president of systems engineering at Smart Vision Works. âYou need a lot of pictures of a lot of carp. “
Ultimately, Smart Vision Works created a hatch design with a fence and a camera. As the fish passed, they were identified and sorted by the exit gates.
MORE FROM STATETECH: How can machine learning secure IoT data?
Computer vision applications: drone mapping and monitoring
Levatas provides smart drone cameras and fixed locations for oil and gas, telecommunications and power generation companies. He currently works with Florida’s largest utility.
âWe help organizations a lot to automate manual inspection tasks,â says Daniel Bruce, partner and product manager at Levatas. These include repetitive and time-consuming safety checks that take up a lot of plant workers’ time. Drones are programmed to detect leaks and other anomalies in hard-to-reach places. Meanwhile, the gauges can be monitored by a fixed smart camera.
âThese are mundane and frustrating jobs,â says Bruce. âWe are not here to replace humans; we are here to increase them and give them superpowers.
The Federal Aviation Administration prohibits drones from flying beyond the operator’s line of sight. Fortem Technologies has created its SkyDome platform to provide autonomous 3D surveillance of airspace. It uses AI-assisted radar to monitor overhead obstacles when a municipality wants to send a drone over the horizon to investigate an accident or crime scene. The system then signals an open flight path to the operator and the FAA provides a waiver.
âYou need a service to tell you that the road is clear,â says Timothy Bean, CEO of Fortem. âWe provide this kind of airspace data so that people can perform a myriad of autonomous drone operations. The company is currently working with 14 state transportation departments.
Edathil of the Texas Department of Information Resources spends his days working on new applications for machine learning in state agencies. A recent proof of concept app used biometric AI facial recognition to register visitors at the front desk.
âThe main advantage of facial recognition is that it minimizes contact with the user,â says Edathil. âThe system has been configured to delete the image after logging outâ, so it does not store personal identifiers without authorization.
Edathil envisions many other uses of intelligent AI in state public agencies. These could include tasks intended for the public at the Texas Parks and Wildlife Department, building inspections by the Texas Facilities Commission, or a toll system that detects expired tags and extracts information from registration stickers.
âIt could also be used to help law enforcement,â he says. “The system can be trained to detect whether the person in the driver’s seat is an authorized driver and potentially report the vehicle theft.”