How can I find depth of an object on a table using Kinect V2 sensor?

I am working on a project where I need to detect an object using Kinect V2, find its 3D coordinates then send them to a robot via TCP/IP. The robot will then move to the object pick it up and place it to another place.

I have done the image processing in Matlab using color segmentation, and found the centroids (x, y) of the object, done the tcp/ip communication and coordinates received in the robot. My problem is on the depth of the object which I must find from the Kinect. How can I be able to find the depth of the object using Kinect, so that I send 3 coordinates to the robot.

Thanks in advance.

Answer 1

[Priysha Aggarwal]

You can easily find depth of an object using concept of Stereo Depth Estimation.

Another pdf with a nice explanation of the concept is here.

Enjoy !

[Jonathan]

I cannot quite understand the Stereo Depth Estimation that you mentioned as to how I can merge it to Kinect V2 sensor. Can you help me on that or with a sample code on how I can go about it in Kinect.

Thanks