Smarter Cameras – What Will You Build?
Cameras have moved beyond a simple device for capturing family vacation pics. Even in industrial uses, the reading of barcodes and labels seem like child’s play when compared to the machine vision applications in use today. As depicted below, smart cameras are being built into a range of devices from the mobile phone and head mounted displays (HMDs), to robotics and autos, and in an ever-increasing number of other ‘things’. For developers, these device options create a new opportunity to build a wide range of applications and experiences.
What Makes A Camera Smart?
A simple camera is an eye that records information, while a smart camera includes a brain that can help it process that information. The standard components today that make up the brain of a smart camera include: image processors, image sensors, frame grabbers, internal memory, lighting and trigger controllers, internet connectivity, GPS, and remote control.
What Makes A Camera Smarter?
Combining a powerful eye and brain with sophisticated software and powerful processors makes an even smarter camera. When combined with existing deep learning technology these cameras can decipher color, read labels, pinpoint scene changes and identify features such as faces.
At Qualcomm Technologies, we’ve added three modules to our camera setup:
Passive depth sensing
Active depth sensing
A combination of hardware with this powerful software is designed to allow for the calculation of contours, angles, slopes and thickness to within less than a millimeter of accuracy opening the door for a growing array of applications for consumers and enterprise as shown below.
As cameras get smarter we can use them to offload our own visual processing. Anything that humans can do with our eyes machines will soon be able to do as well.
Whatever your role in computer vision and camera development, Qualcomm Developer Network has the tech, tools, and resources to help. Sign up for QDN's email updates to keep up with the latest.