FULL TUTORIAL TO CREATE YOUR OWN HAAR CASCADE: https://www.youtube.com/watch?v=ydSXgBZ1ybk
UPDATE [7-12-2018] I've also built an UI for it in Tkinter: https://github.com/AdityaPai2398/UI-For-Vehicle-and-Pedestrian-Detection-/tree/master/Vehicle%20and%20pedestrain%20detection
Computer Vision Plays a vital role in traffic management and surveillance systems and has been an active research area in the past years. In systems like these, the detection of vehicles and also classification of the vehicle plays a major role. The datasets are traffic videos of urban environment taken from various cities around the world which were used to train the classifier hence generating a robust classifier. The proposed approach is computationally less expensive with faster processing speed. The experiments on-road prove it to be a robust and real time algorithm which is highly competitive with the existing architecture.
This System has the following Haar Cascades:
- Haar Cascade for Car detection
- Haar Cascade for Bus Detection
- Haar Cascade for Two-Wheeler Detection
- Haar Cascade for Pedestrian Detection
The Results of the system are as shown below:
Steps To Run The Project
System Requirements:
- Python 3.6 Or above
- Opencv-python -> Python 3.4+ comes with an easy installation tool called pip to install additional packages. (package manager)
Successfull installation of above packages may be checked by moving into python shell and typing "import cv2" If nothing shows up then installation was successful. . We had used python 3.6.1 but any version higher than 3.6.1 will also do.
The Project Folders have 4 sub folders
1)Car
2)Two Wheeler
3)Bus
4)Pedestrian
In each of the subfolder there will be two videos ,one haarcascade ,two python scripts (one for each of the videos).
Due to file size constraint I wasn't able to upload all the videos. Please raise an issue if you need the videos and I'll send it ASAP.