Self Driving cars with Opencv
summary
A software prototype built with OpenCV detects traffic lights (color thresholding, white top-hat filter, blob shape filtering), road signs (HSV + Gaussian + Canny + contour/ellipse), pedestrians (pretrained HOG + linear SVM) and lanes (grayscale, blur, Canny, mask, Hough transform). It estimates obstacle depth and time-in-frame, classifies moving vs stationary, and issues real‑time motion decisions as a proof‑of‑concept (software only, no hardware steering yet) while calling out real-world failures around pedestrian intent and emergency handoffs.
AI-written summary · about
chapters (11)
- 0:00Introduction to self-driving cars
- 1:12What is a self-driving car
- 2:15History and challenges of self-driving cars
- 7:13Role of OpenCV in self-driving cars
- 9:21Building the primary prototype
- 10:02What is OpenCV and why use it
- 11:08Traffic light detection methods
- 17:03Traffic sign detection techniques
- 21:15Pedestrian detection with OpenCV
- 23:02Lane detection process
- 25:11Conclusion and Q&A
related talks
46mDriver Assistant Solutions with
49mBuilding a self driving RC car
28mMachine Learning Inside the Car
22mOpenCV and TensorFlow in practice
29mRecognizing Traffic Lights With Deep Learning
41mAI applications in autonomous driving
31mAutonomous driving, how Python drives Formula Racing cars
47mAn unexpected journey into computer vision, or There and Back Again