Simple ADAS Project

In this project, We developed ADAS(Advanced Driver Assistance System) by using C++ language.

Inspired by

• https://github.com/ultralytics/yolov3

Project environment

• Ubuntu 20.04.3 LTS
• OpenCV 4.5.5
• GeForce GTX 1060 6GB
• CMake 3.16.3
• Used USB video camera (resolution: 640 * 480)

Test video link

• https://youtu.be/DmDQBkF9rJg

Project mechanism

Lane detection

1. Convert video's color space from BGR to HSV.
2. In HSV Color space, by using cv::inRange(), we extract the images that have the same color as the lane.
3. Pass the low-pass filter to remove noise.
4. Do dilation operation to make lane more clear in the image and then extract the lane by using Canny edge and HoughLines Transform. 5.Lane departure warning alert when the driving car escape the detected lane.

Object detection

1. Used yoloV3-tiny pre-trained weight (COCO) for car & other object detection.
2. If the detected object's location is located within the pre-specified ROI(In front of the car), collision dager alert perform.

Features

• This program shows us whether the driver keeps his lane and a safe distance.
• Upper left side: Print Safe if the driver stay in his lane. Print Warning if it's not.
• Upper right side: Print Safe if the driver keep a safe distance. Print Warning if it's not.