mapping coordinates from specific point in point cloud using image pixels

i am using yolo detector to detect strawberrirs the detector draws a bounding box around the strawberry, the process is done using kinect v2 camera, with this camera i am able to getting RGB image and depth image with these two i can creat point cloud, so what i need is using the center of the bounding box (which is in pixels) and map it to the point cloud so i get XYZ in real world, because i want to send these valuse to a robot.

[KSSV]

Read about imcrop

[Zakariya Abousabie]

hi @KSSV how can i use it in my problem as i read about it it just crop images to size you want in my problem i have image that i creat point cloud from it, what i need is to use image pixels such as I(200,350) to know the x y z crosponding to these pixles

[Zakariya Abousabie]

@Walter Roberson @Image Analyst hey guys i have seen that you are one of the best people have answers to situations like mine so if you can help me i will be appreciated.

Answer 1

[sudobash]

Hi!

As per my understanding, the world coordinates need to be computed from a given point in an image. You can use pointsToWorld function to do this. The documentation for the same explains this in detail. All you would need are the intrinsic and extrinsic parameters of the sensor.

Hope this helps!