driverless-eyebot

Note:This project is based on eyebot system develop by The Univ. of Western Australia, so if you want to use it you should replace the eyebot API code(controlling the motor) by your own function.

Carolo-Cup is an international autonomous driving competition held in German annually. Student teams need to compete with each other about the new ideas and actual performance of their driverless model vehicles. The test platform built by Technische Universität Braunschweig contains different scenarios, so it can well simulate the real-life situation. In other words, although this is a model, the experience gained from solving problems in the model world can be very helpful to the realistic environment and this is the most fascinating part of this game. This project has done the lane detection, traffic sign detection and classification and communication between eyebots and base station.

Preparation for building

make sure you have installed cmake.
download dlib from dlib, and put it in your root folder.
make sure you have installed opencv.
(this step is for eyebot) replace the /home/pi/eyebot/bin/sound as the sound.
(this step is for eyebot) copy eyebot++.h&eyebot.h&types.h in /home/pi/eyebot/include into /usr/include and change their permission by sudo chmod 777 eyebot++.h eyebot.h types.h.
(this step is for simulation) download cygwin from cygwin (pw:5uiz), this version includes eyebot lib and opencv lib.

How to build

stand in your root folder
run mkdir build
run sh make
if you want to add new source file add it in CMakeLists. And the source file should be included in src folder.
the executing file will be stored in the build folder, the final version is called eyebot.

Traffic sign detection part

This part can run on PCs. Building is same as above. And if you just want to run the traffic sign detection part please comments link_directories("/home/pi/eyebot/lib") and target_link_libraries(${name} libeyebot.a) in CMakeLists.

Mark signs on traffic sign images

Compile imglab:

cd dlib/tools/imglab
mkdir build
cd build
cmake ..
cmake --build .

Create XML from sample images:

dlib/tools/imglab/build/imglab -c images/pare/training.xml images/pare/train/*.(jpg|jpeg|png)
dlib/tools/imglab/build/imglab images/pare/training.xml
dlib/tools/imglab/build/imglab -c images/pare/testing.xml images/pare/test/*.(jpg|jpeg|png)
dlib/tools/imglab/build/imglab images/pare/testing.xml

Use shift+click to draw a box around signs.

Train the fHOG detector

To train a fHOG detector, run build/hog_detector. For example, to run the detector on the image/stop/ folder in the verbose mode, execute the following command:

build/hog_detector -v images/stop/

The detector will be saved to the file detector.svm. To change the file, use the --detector-name option.

Run build/hog_detector -h for more details.

If you want to view traffic sign detection results on computer

use the parameter --wait to wait for user input to show next image.

build/detect --wait exam/*.bmp

lane detection;

traffic sign detection and classification;

using cmake to compile;

Locke637 / driverless-car-eyebot