A technology company that uses computer vision to mine traffic video for data recently released a free cloud platform to which departments of transportation and engineers can upload an unlimited amount of footage and traffic data and find and connect with video consultants and contractors. Additionally, the company is releasing a new piece of hardware that plugs traffic signals into the “internet of things” (IoT).
“My main use for Miovision Central is as a centralized database for all the traffic counts we do,” says Vanessa Jorge, Maser Consulting, Red Bank, N.J. “Traffic counts” track the number and type of vehicles that pass through traffic lights during specified hours of the day. “Before, I’d get traffic counts via email as attachments, and I’d have to do filters and searches through my email to find the data [to] send along to my colleagues,” says Jorge.
Miovision Central also allows Jorge to search for consultants to collect data. Once a consultant is hired, they can use the platform to transfer data back and forth. The database is searchable, so instead of doing keyword searches in email and looking through attachments, it is possible to search by location, date and time.
Further, as in Jorge’s case, a user can switch over to a map view and see all the traffic lights she is monitoring. “Miovision Central is the sharing platform for the traffic data we’ve been sending around for the past decade,” says Cam Davies, senior product marketing manager at Miovision.
His company also released a new piece of hardware called Spectrum, which communicates with traffic signals and connects them to the Miovision platform so that customers can manage the traffic lights in an interactive way.
“The technology that connects intersections is terribly old and inefficient in most locations,” says Davies. “Often, you’re at a light for way too long with no other cars on the road. By using Spectrum, they can know if the lights are failing or when the power drops,” says Davies, adding that the most common way DOTs currently detect failed traffic lights is through citizen complaints.
With this new product, DOT staff can get an email or text message that alerts workers to the problem and its location, says Davies. Miovision began in the traffic data mining sphere by selling a video collection unit called Scout—a traffic signal-mounted video camera that records traffic patterns.
Miovision works primarily as a consultant, using computer vision to mine video footage for data such as the volume of traffic, types of vehicles passing by and their turning direction. “You’ll get data readings such as: 200 cars and 80 trucks turned left between eight and 8:15 a.m.,” says Davies.
The alternative is putting workers on the corner to count the cars, trucks and bikes during a certain time frame. Miovision prices its services to be comparable to what a minimum-wage worker would make doing that job, says Davies.
“Our hourly rate ranges from a buck or two an hour to $20 an hour, depending on how complex the data is,” says Davies. “We process about 20,000 hours of video every day.”
Davies says Miovision data is verifiable by the raw-video files, but, often, a worker’s data can be verified only by tallies in a notebook.