Computer Vision Interview
20 essential Q&A
Updated 2026
stereo
Stereo Vision: 20 Essential Q&A
Disparity, baselines, rectification, and classical dense matching for depth maps.
~11 min read
20 questions
Advanced
disparitybaselinerectifySGBM
Quick Navigation
1
What is stereo vision?
⚡ easy
Answer: Using two (or more) calibrated views with known baseline to recover depth via triangulation of corresponding points.
2
Define disparity.
📊 medium
Answer: Horizontal shift between corresponding pixels in a rectified stereo pair—larger disparity means closer surface (inverse relation to depth).
3
Depth from disparity?
📊 medium
Answer: Z ≈ f × B / d (f focal length, B baseline, d disparity)—assumes rectified parallel cameras and pinhole model.
4
Baseline tradeoff?
📊 medium
Answer: Larger B increases depth precision (more parallax) but worsens occlusions and matching in narrow scenes; small B reduces measurable disparity range.
5
What is rectification?
🔥 hard
Answer: Warp both images so epipolar lines are horizontal scanlines—reduces correspondence search to 1D and simplifies disparity.
6
Epipolar constraint?
📊 medium
Answer: Without rectification, match for a point lies on a line in the other image—comes from epipolar geometry of two views.
7
What is stereo matching?
📊 medium
Answer: For each pixel (or patch), find best match along epipolar line using photometric cost (SAD, census, CNN features).
8
Cost volume?
🔥 hard
Answer: 3D array H×W×D of matching costs over disparity levels—winner-take-all or global optimization (SGC, belief propagation) picks disparities.
9
What is SGM?
🔥 hard
Answer: Semi-Global Matching aggregates costs along many paths with smoothness penalties—good quality/speed tradeoff in OpenCV StereoSGBM.
10
Occlusion regions?
📊 medium
Answer: Pixels visible in only one view have undefined disparity—detected by consistency checks or left-right validation.
11
Sub-pixel disparity?
📊 medium
Answer: Parabolic fit around discrete minimum or phase-based methods—needed for smooth surfaces and accurate 3D.
12
Common errors?
📊 medium
Answer: Calibration errors, textureless regions, repetitive patterns, specular highlights, and motion if scene moves between exposures.
sgm = cv2.StereoSGBM_create(minDisparity=0, numDisparities=128, blockSize=5)
disp = sgm.compute(imgL, imgR)
13
StereoBM vs SGBM?
⚡ easy
Answer: BM: fixed small block, fast, blocky. SGBM: semi-global, slower, smoother—preferred when quality matters.
14
Monocular depth?
📊 medium
Answer: Single image lacks scale without priors—learned networks predict relative depth; stereo gives metric depth with calibration.
15
vs RGB-D?
⚡ easy
Answer: Structured light / ToF gives depth directly—no correspondence problem but range/resolution limits; stereo passive but needs texture.
16
Multi-view stereo?
🔥 hard
Answer: Fuse many images (MVS) for dense point clouds—used in photogrammetry beyond two-camera stereo.
17
Stereo in driving?
📊 medium
Answer: Wide-baseline camera pairs on vehicles for obstacle depth; often fused with radar/LiDAR and learned refinement.
18
Fuse with LiDAR?
🔥 hard
Answer: Sparse accurate LiDAR anchors depth map from stereo; learning-based fusion common in autonomy stacks.
19
Learned stereo?
📊 medium
Answer: CNNs build cost volumes or regress disparity directly (e.g. PSMNet)—strong on benchmarks when enough training data.
20
Need calibration?
⚡ easy
Answer: Yes for metric depth—need K, distortion, and stereo extrinsics; rectification matrices derived from them.
Stereo Cheat Sheet
Geometry
- Z = fB/d
- Rectify
Match
- Cost + smooth
- SGM
Issues
- Textureless
- Occlusion
💡 Pro tip: Rectify first so disparity is a 1D search per row.
Full tutorial track
Go deeper with the matching tutorial chapter and code examples.