If you have ever seen a Google Street View car gliding around with its camera rig on top taking pictures of your area, your first inclination might be to spam the cameras by doing something silly. I mean you certainly wouldn't be the first person to point at your junk or play dead for the cameras. Google has reportedly upgraded the Street View camera fleet to gather more information to feed its machine learning and AI tech back at home base.
According to a report from Wired, this is the first significant update to the tech used by the Street View fleet in eight years. Previously, the camera towers mounted to the roof of the car housed 15 cameras to capture footage. The new Street View camera system has only seven cameras inside and each of those cameras has a 20MP sensor for high-resolution photos.
The new roof-mounted system also has pair of HD cameras for still shots and those round cans on the front and back are said to hide laser radar. All of the raw data is fed through algorithms so that important information like street numbers and street names can be added automatically to the Google database.
The software Google uses is said to be smart enough that it can recognize business logos and names, making it easier to find that local Starbucks or McDonald's. Google is also working on improving the system so that its algorithms can read data from smaller signs to automatically gather business hours. Google's VP of its mapping branch Jen Fitzpatrick said that the goal is to allow Google Maps to answer questions that reference how the world looks.
She says that one day Google Maps will be able to answer questions like "What’s the name of the pink store next to the church on the corner?" Fitzpatrick says that in order to answer that sort of query, Google needs deeper and richer information. There is no clear indication of what the laser radar is doing; perhaps it is gathering detailed data on roads to help with Google's autonomous driving efforts.
Image via Wired