Many large U.S. cities seem to have adopted the motto: “Man, parking here sucks.” But a Swedish startup believes it can apply a little technology to solve a problem that is costing billions of dollars each year.
Today, Mapillary, based in Malmo, Sweden, announced it will collaborate with Amazon Rekognition, a visual data analysis platform, to create a system that makes it easier to find parking in congested areas.
Mapillary is developing a street level-imagery platform that uses computer vision technology to enhance digital maps. The company will now use Rekognition’s text-in-image feature to scan parking sign data on a large scale.
“City authorities are struggling to keep track of their parking signs and data,” said Mapillary CEO and cofounder Jan Erik Solem, in a statement. “For cities to manually track this themselves would cost millions of taxpayers’ dollars, not to mention the enormous time investment. The collaboration between Mapillary and Amazon Rekognition means that several cities in the U.S. are now in a position to get complete updates on their parking signs through an automated and computer vision-driven process, saving both time and money.”
Mapillary says it currently has 350 million images on its platform. These images are collected via contributions from various people and organizations. While Mapillary can detect whether an image is a parking sign, it can’t actually read the text. Amazon’s Text-in-Image Rekognition will make that possible.
Mapillary says tests of the system using 40,000 parking signs resulted in the text being read correctly 95 percent of…