Car and lane detection with opencv and yolov3 c++
This is a method that detects cars and lane on road using c++ opencv and yolov3. These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
System version that I am using
opencv => 2.4.9
gcc => 5.4
yolov3
Hardware
Intel® Core™ i5-7500 CPU @ 3.40GHz × 4
8gb Ram
Installing
git clone https://github.com/lamprospapav/Car-and-lane-detection-with-opencv-and-yolo-cpp.git
cd Car-and-lane-detection-with-opencv-and-yolo-cpp/
For yolo detection you need download yolo3.weights from wget https://pjreddie.com/media/files/yolov3.weights and copy it at YoloNet folder. In the next version i will train my own weights for car detection.
mkdir build
cd build
cmake ..
make
cd ..
cmake .
make
./lane_detection
Running the tests
For your camera you need to calibrate it using your chessbord images into camera_cal directory.
I am running code with cpu and I need 80 ms for a frame. Input image yolo net : 224 x 224. In next version i will use darknet and gpu.
output video : https://1drv.ms/v/s!Anr_pcsiuqDUj7pK2mk8_SxKhQzafg