Yousuf Rafi
High-quality camera means high-quality photos and feed more data for Google’s machine learning capabilities. The advanced camera design is the major upgrade in the past eight years that can produce high quality Street View imagery and help AI to index real-world locations.
The trend of huge, globe-like cameras on top of Google Street View cars is gradually dying. Google has updated the design for the cameras used to capture the images for Street View for the first time in the past eight years. The new camera design is simple, the rig has the potential to capture photos that are clear, higher in resolution and brighter in color.
Related: Google Algorithm just got updated
The rig also serves as a host to two cameras that can take sharper HD photos and two cans on the front and back for laser radar. These cameras can take clearer, closer and brighter shots that make it easier for Google’s image recognition algorithms to identify buildings and street signs.
These algorithms can store millions of signs, business names and storefronts. Google also aims to improve its already challenging digital mapping database. The basic idea behind this Street View camera advancement is to enhance machine learning and AI capabilities. When pictures are taken by a Street View car, algorithms can easily detect and note relevant street names and numbers automatically adding them to Google database.
The software also has a potential to identify business names and brand logos. Google is still working to further improve the software and enhancing its capabilities to make it enable to determine different types of stores based on what they look like, and reading smaller details like when the store opens and closes.
This advanced technology will provide more accurate data that will boost Google’s machine learning capabilities and ultimately expand Google’s search results and Virtual Assistant functionalities.
In the past few years, Street View cars have captured more than 80 billion photos in thousands of cities and 85 countries. Google’s traditional mapping data is even more extensive. But Google still wants better index of the world.
Jen Fitzpatrick, Google’s Vice President and Head of Map division says, “Google Maps might be able to provide more detailed results because the company is getting deeper and difficult questions. And harder questions can only be answered if we have deeper, advanced and more information. To answer such tougher queries, we need a fresher and detailed digital model of the word.”
She wants her service to deal with queries such as “What’s the name of the brown store next to the convenience mart on the corner?”
Google’s biggest investment in AI and machine learning offers a natural way to get accurate information. Thanks to recent research inside the maps division, when a Street View car takes images of a road, algorithms can automatically create new addresses in the map’s database by locating and transcribing any street names and numbers.
High quality images captured from the new cameras on Google’s Street View vehicles will allow algorithms to extract information more reliably.
“From a machine learning perspective, everything gets better.”, says Andrew Lookingbill, technology engineer. It will also help teams to develop advanced software.
Google Street View has been functioning since 2007, and the fleet of Street View vehicles includes cars, snow mobiles, trikes and a trolley that takes photos indoor in places like museums. The camera rig has been outfitted onto a backpack to capture photos of places that aren’t easily reached by vehicle.
Street View imagery with algorithms can particularly be useful in places where cities, businesses and roads are updating fast. Google’s image sourcing algorithms can help translate the new imagery into a significant bump in map quality. Moreover, Google promotes digital ad industry by selling ads inside maps, so new coverage and authenticity can lead to more profits if a company draws new users and usage to the service.
Google wants people to feed the image-hungry algorithms. The tech industry recent updates in virtual reality has made immersive 360-degree camera relatively cheap. Google has also started certifying cameras called Street View ready that allows you to upload your own footage throughout the Street View mobile app to live on the company’s service. The panoramas will be processed by Google’s image recognition algorithms for fresh map data just like its own imagery data.
This advancement in camera setup will improve results in Maps, help VR with clearer imagery and self-driving car platform.
Ashad Rehman
