Posted by Anirudh Bhardwaj | Last Updated: 07-Mar-17
Artificial Intelligence has bridged the gap between machines and humans by emphasizing some astonishing aspects and some groundbreaking technologies like Machine Learning, Computer Vision and Image Processing. It might never have occurred to you but your day-to-day life is very much influenced by Artificial Intelligence. For instance, if you’re frequently visiting an e-commerce site, it automatically keeps a track of your search preferences. So the next time you visit that site, you’ll start seeing results based on your search preferences. This technique of learning user behavior is called pattern recognition which is an integral part of Machine Learning and Artificial Intelligence.
Computer Vision and Robot Vision are the two most astounding inventions in the field of AI but as it turns out, there’s quite a lot of misconception about these two terms or entities. More often than usual, the two entities are perceived to be the same. But the thing is, there are numerous major differences between the two. While the two entities can never be the same, it is a common mistake that many people make owing to the striking similarities between Computer Vision and Robot Vision. In this blog, you’ll know how these two technologies are different from one another and where do we have to draw the line.
You may also like New Computer Vision Technology To Identify Distracted Drivers.
What Is Robot Vision?
Robot Vision as the name suggests, is a set of algorithms which renders vision to the robotic components. At its core, Robot Vision is a combination of computer algorithms, cameras and other hardware components that work unanimously in order to provide visual insights to the robot or machine. This helps the robot to accomplish complex tasks that require visual understanding.
For instance, if it wasn’t for the Robot Vision techniques, a machine or say a robotic arm won’t be able to pick an object and place it somewhere else if need be. You can imagine this scenario where sensors and cameras are used to detect an object placed on a wooden plank which is then lifted up by the robotic arm using complex Robot Vision algorithms. For object detection as discussed in above mentioned case, the Robot is equipped with the ordinary 2D cameras while this can be problematic if it were to mount its wheels on a moving vehicle. In that case, you’ll need advanced 3D Stereo Cameras instead.
So that’s what Robot Vision is all about and seeing its use, one can easily conclude that a Robot would apparently be blind if it wasn’t for Robot Vision.
Computer Vision is a subfield of Artificial Intelligence which aims on rendering advanced visual capabilities to the computers by means of complex algorithms and camera hardware. It mainly deals with image recognition. The computer vision methods initially extract useful information from the digital images and videos. This information is then processed and analysed. Finally with help of cameras (usually webcam), the computer acquires high-dimensional data from the real world. This data is further processed to generate symbolic information which enables the computer to take important decisions.
There are several other subfields of Computer Vision including object recognition, video tracking, object pose estimation, image restoration and event detection.
Robot Vision vs Machine Vision
While the above explanation contrasts between Computer Vision and Robot Vision, sometimes people still relate Robot Vision with Machine Vision. The two terms are apparently the same but there are certain points where these two contradict. Machine Vision mainly finds its use in industrial domain. Some notable applications of Machine Vision are automatic inspection, robot guidance and process control. But it must be taken into account that although Machine Vision methods are used to guide robots but it’s still different from Robot Vision. First off, Machine Vision is comparatively a broader field and many of the Machine Vision applications have nothing to do with Robotics. On contrary to this, Robot Vision is primarily meant to be implemented inside the robotic landscape alone and that’s where we draw the line.