What is the principal difference between 2D vision and 3D vision in robotics?

• A. 2D vision uses image features for detection; 3D vision provides depth/pose using stereo, structured light, or depth cameras.
• B. 2D vision provides depth; 3D vision provides color only.
• C. 2D vision uses depth cameras; 3D vision uses 2D sensors.
• D. 2D vision is only for color images; 3D vision is only for infrared

Answer: (A)

In robotics, the main distinction is that 2D vision gives information in the image plane: you detect features, edges, textures, and colors to recognize or track objects, but you don’t directly know how far away things are. Depth and real-world geometry have to be inferred indirectly, which can be challenging for precise tasks like grasping or accurate navigation.

3D vision adds depth information and the ability to estimate an object's pose in space. This comes from techniques like stereo vision (using two cameras to compute depth from disparity), structured light (projecting a pattern and reconstructing shape), or depth cameras (time-of-flight or structured-light sensors). With depth data, you get a 3D representation (point cloud, depth map) and can determine the exact position and orientation of objects, enabling more reliable manipulation and spatial reasoning.

So the best description is that 2D vision focuses on image-based detection and features in the picture, while 3D vision provides depth and pose information through stereo, structured light, or depth sensors. The other statements mix up capabilities—for example, 2D is not inherently depth-producing, and 3D is not limited to color-only or infrared—since 2D can use color or grayscale, and 3D sensing spans several modalities.