Build Your own Self Driving Car | Deep Learning, OpenCV, C++

This lecture outlines hardware essentials for a self-driving car: chassis kit, four dc gear motors with tires, L298 driver, Arduino, Raspberry Pi 3 model B plus, camera, and power source.

Assemble the robot chassis by mounting two dc gear motors and wiring them in parallel across motor bridges, cross-coupling wires to ensure clockwise rotation.

Build a testing track for a self-driving car by gluing chart segments into a path, adding curves and straight sections, and using white strips for lane guidance in image processing.

Learn to control DC geared motors with Arduino by implementing forward and backward functions, configuring pins, using PWM on enable pins, and testing the hardware.

Develop left and right motor controls by adapting the forward function, lowering speeds to steer, and create three left-turn levels and three right-turn levels tested via Arduino.

Flash the raspbian os onto a raspberry pi 3 b+ by preparing a 16 gb sd card, formatting it, downloading the raspbian stretch desktop image, and flashing with a tool.

Connect the Raspberry Pi to a PC via Ethernet, enable SSH on boot, locate its IP with an advanced IP scanner, and access the GUI via remote desktop.

Connect a Raspberry Pi to a personal computer over wifi with no internet cable, power via usb, and use IP scanner to find its IP for a remote desktop connection.

Connect a Raspberry Pi to your personal computer using a VNC viewer, enable the VNC server, and compare VNC performance with remote desktop options.

Learn to restore a disk image from an SD card backup to a Raspberry Pi, format partitions, and boot a preinstalled OpenCV environment for a self-driving car project.

Learn how to create and restore my SD card backup to set up a Raspberry Pi project, install OpenCV, and run prebuilt robotics software.

Explore OpenCV for computer vision and image processing, install it on a Raspberry Pi, and use C++ bindings to build cross-platform machine learning apps with cascade classifiers.

Prepare a Raspberry Pi for open tv installation by adding an exhaust fan, monitoring temperature, connecting via Wi-Fi or ethernet, and removing unwanted software while expanding storage.

Clone OpenCV from GitHub and set up a Raspberry Pi development environment, update the OS, configure storage, and install cmake, pkg-config, and essential libraries for OpenCV.

Build OpenCV on Raspberry Pi using CMake by downloading from GitHub, creating build directories, and compiling with make to enable self-driving car workflows.

Set up and link OpenCV libraries in the programming editor by creating a package config file, copying a template, adjusting permissions, and organizing the local libraries folder.

Create your first Geany project, save as C++ file, and write an OpenCV test program to verify installation. Configure the build to link OpenCV via package config, fix errors, build.

Install and configure the raspy cam library on raspberry pi, clone from GitHub, build and install, configure libraries and headers, and verify OpenCV and wiringPi integration for camera access.

Mount a Raspberry Pi and camera on a self-driving car chassis, place the camera at the back with height using wooden sticks, and tilt slightly downward for better track visibility.

Back up your Raspberry Pi SD card to preserve installed OpenCV and Raspbian libraries, then restore the backup anytime using disk management software.

Learn to initialize and access the Raspberry Pi camera in a C++ OpenCV project by enabling the camera interface, rebooting, and writing code to connect, verify, and run.

Learn to capture images with c++ using the raspy cam library by configuring a camera object, including resolution, brightness, contrast, saturation, and frame rate.

Capture video from a raspy cam in C++ by looping grab and retrieve of frames, using 360 by 240 resolution for image processing and bitrate adjustments to reduce lag.

Compute frames per second for a video stream in C++ using the chrono library to measure elapsed time from start to end, convert to seconds, and display the fps.

Explore image processing with OpenCV by converting frames from BGR to RGB, capturing video on a Raspberry Pi, and comparing BGR and RGB displays to reveal the color signature.

Create a region of interest for track processing with OpenCV by selecting the bottom frame half, defining four points, and applying a perspective warp to a bird's-eye view.

Map a region of interest to a bird's-eye view by defining four source points and four destination points, applying get perspective transformation and warp perspective in OpenCV C++.

Convert frame perspective to grayscale, then apply inRange threshold to extract lane lines, adjust thresholds for outdoor vs indoor conditions, and reduce noise for clearer results.

Implement canny edge detection on grayscale images, tune threshold and hysteresis, and merge edge frames to produce a final visualization for a self-driving car vision system.

Troubleshoot hardware and software for a self-driving car by adjusting camera position, frame width, and region of interest to achieve a wide-angle view and reliable lane detection with edge detection.

Identify lane lines by defining a region of interest and computing their left side positions using a rectangular ROI, histograms, and per strip intensity analysis in OpenCV.

Create a histogram function using a dynamic vector to store grayscale intensities from a region of interest; in a 400-width frame, iterate, normalize to 0–255, and push values.

Locate lane lines by using iterators to find maximum intensity in a dynamic image and compute positions with distance, then draw green lines on the frame.

Compute the lane center from the left and right lane positions, then calibrate by shifting the frame center to overlap with the lane center.

Compute the lane-center offset and steer the self-driving car between lanes by displaying the offset on frames and preparing Raspberry Pi to Arduino communication.

connect a raspberry pi to motor controls using digital pins and pin mode setup. implement conditional logic to drive forward, left, or right based on sensor-like values.

Install and fix the Wiring Pi library on Raspberry Pi by uninstalling the existing version, installing the latest core, building the library, and updating the package configuration.

Program the Arduino Uno as a slave device by reading digital pins 0–3 from a Raspberry Pi, convert binary to decimal, and drive motors with forward, right, left, and stop.

Test and tune your self-driving car by compiling the program, adjusting left and right motor presets, and verifying center alignment on the external camera track.

Decrease the data load by switching from 32-bit color to 15-bit color, reducing frame delay when streaming from Raspberry Pi to the PC over a remote desktop connection.

Test the self-driving car on a large track with 12–15 black darts; place the car, run the program, and adjust the reason of interest for new conditions.

Use a histogram of intensity values to detect the lane's white strip, updating a dynamic editor and issuing a U-turn or move commands via Arduino and console.

Implement and test a U-turn function for the slave device in a self-driving car project, including camera stabilization delays and left-turn motor sequencing at the lane end.