Introduction

vehicle recognition and tracking based on yolo v3 and sort algorithem.

This is the final project of the machine vision course. Given that multi-object tracking (MOT) is currently the research focus, this project will open source to help related research. The basis of this project is: YOLOv3, multi-target recognition algorithm, Kalman filter tracking, Hungarian algorithm. (IPIL 2016) This project has basically completed the identification and tracking of vehicle targets in the video, but the robustness is not perfect.The yolov3 implementation is from darknet. BTW, this project can work well in online network.

To run this project you need:

• python 3.6
• numpy
• scipy
• opencv-python
• sklearn
• pytorch 0.4 or 1.x
• time
• filterpy
• matplotlib

all of these is load on the request.txt.

Quick start:

1. Clone this file

$ git clone https://github.com/Github-chenyang/vehicle_tracking_yolov3.git

2. You are supposed to install some dependencies before getting out hands with these codes.

$ cd vehicle_tracking_yolov3
$ pip install -r ./docs/request.txt

for user in china, you can specify pypi source to accelerate install like:

pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple 

3. Download YOLOv3 parameters

$ cd config/
$ wget https://pjreddie.com/media/files/yolov3.weights
$ cd ..

4. Run demo just run the object_tracker.py and you can realize a simple offline MOA. here is the effect of the algorithm

train your own cfg as well as weights

This part is mainly based on packyan’s work. I just change some code so that it is easy for me to use.

Download pretrained weights

if you wan use pretrained darknet-53 on IMAGENET weights, please download darknet53.conv.74,and put it into checkpoints/

if you just want a pretrained weights on kitti dataset for test or detect, please download pretrained weights file, and put it into weights folder, the path: weights/yolov3-kitti.weights

if you want some proper vehicle dataset:

I truly recommend the NVIDIA city dataset.Data for this challenge comes from multiple traffic cameras from a city in the United States as well as from state highways in Iowa. maybe this is its looklike.

final

the project has been updated and the performance is better now! just enjoy!