There are different computer vision problems you need to solve in a self-driving car.
βͺοΈ Object detection
βͺοΈ Lane detection
βͺοΈ Drivable space detection
βͺοΈ Semantic segmentation
βͺοΈ Depth estimation
βͺοΈ Visual odometry
Details π
Object Detection ππΆββοΈπ¦π
One of the most fundamental tasks - we need to know where other cars and people are, what signs, traffic lights and road markings need to be considered. Objects are identified by 2D or 3D bounding boxes.
Relevant methods: R-CNN, Fast(er) R-CNN, YOLO
Distance Estimation π
After you know what objects are present and where they are in the image, you need to know where they are in the 3D world.
Since the camera is a 2D sensor you need to first estimate the distance to the objects.
Relevant methods: Kalman Filter, Deep SORT
Lane Detection π£οΈ
Another critical information the car needs to know is where the lane boundaries are. You need to detect not only lane markings, but also curbs, grass edges etc.
There are different methods to do that - from traditional edge detection based methods to CNNs.
Driving Path Prediction ‴οΈ
An alternative is to train a neural network that will directly output the trajectory that the car needs to drive. This can be used as a substitute to centering between the lane markings if they are not visible for example.
Drivable Space Detection βοΈ
The goal here is to detect which parts of the image represent the space where the car can physically drive onto.
The methods here are usually very similar to the semantic segmentation methods (see below).
Semantic Segmentation π¨
Not all parts of the image can be described by a bounding box or a lane model, e.g. trees, buildings, the sky. Semantic segmentation methods classify each pixel in the image.
The goal is to estimate the distance to every pixel in the image, in order to have a better 3D model of the surrounding.
Methods like stereo and structure-from-motion are now being replaces by self-supervised deep learning models working on single images.
Visual Odometry π₯
While we know the movement of the car from the wheel sensors and IMU, determining the actual movement in the camera can be more accurate to get the pitch angle for example.
The visual odometry estimates the 6 DoF movement of the camera between two frames.
Summary π
There are of course many other computer vision problems that may be helpful, but this thread will give you an overview of the most important ones.
As you see, nowadays, deep learning methods (and especially CNNs) dominate all aspects of computer vision...
If you liked this thread and want to read more about self-driving cars and machine learning follow me @haltakov!
I have many more threads like this planned π
β’ β’ β’
Missing some Tweet in this thread? You can try to
force a refresh
Interesting results from the small experiment... π
This was actually a study reported in a Nature paper. Most people offer additive solutions (adding bricks) instead of substractive solutions (removing the pillar).
In this example, the most elegant solution is to remove the pillar completely and let the roof lie on the block. It will be simpler, more stable and won't cost anything.
Some people quickly dismiss this option assuming this is not allowed, but it actualy is π
This isn't because people don't recognize the value, but because many don't consider the substractive solution at all. Me included πββοΈ
The paper shows that this happens a lot in real life, especially in regulation. People tend to add new rules, instead of removing old ones.
End-to-end approach to self-driving π₯ πΈοΈ πΉοΈ
I recently wrote about the classical software architecture for a self-driving car. The end-to-end approach is an interesting alternative.
The idea is to go directly from images to the control commands.
Let me tell you more... π
This approach is actually very old, dating back to 1989 and the ALVINN model by CMU. It is a 3-layer neural network using camera images and a laser range finder.
A modern example is Nvidia's PilotNet - a Convolutional Neural Network with 250M parameters, which takes as input the raw camera image and predicts directly the steering angle of the car.
No explicit lane boundary or freespace detection needed!
Open-Source Self-Driving Car Simulators πΉοΈ π
You want to play around with self-driving car software and gather some experience? Check out these open-source self-driving car simulators!
Details below π
CARLA
CARLA is a great software developed by Intel. You can use it to work on any step of the pipeline, model different sensors, maps, traffic. It also integrates with ROS.
Another great simulator by Voyage - the self-driving company that was recently acuired by Cruise. It is built on the Unreal Engine and supports lots of features.
Useful online courses on self-driving cars π§ π
Here is a list of useful courses if you want to learn about software for self-driving cars.
Some of the courses are paid, but all platforms offer regular discounts and financial aids if you can't affor them.
Thread π
Udacity Self-Driving Car Nanodegree
This program offers hands on experience on all kind of relevant topics like perception, localization, planning and control. It takes a lot of time, but it is worth it.