using metric_depth to read frames from webcam and show point cloud using open3D

[damavand1]

Hi

I want open web cam, detect depth and show mesh or point cloud on screen using open3D, something like when we using depth cameras.

somethings like this for every frame: Link

can you please give me a basic code?

another question is, do you think is it accurate enough to use in a self driving robot like Coco delivery and starship ?

I want to use YoloV8 to segmentation (find road, people, cars and ...) and using Depth anything metric-depth to find where is it in 3d space and finally create a vector space like Tesla (in this step just using single front camera not 8 camera)

thank you

[Kaiz0506]

Hi may I know have you solve it. Thanks!

[mina-start]

Hi, have you successfully calculated the position of an object in space using metric depth? Thanks!