Skip to content

opencvfun/Object-Detection-and-Action-Recognition-for-Driverless-Vehicles

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Object Detection and Action Recognition for Driverless Vehicles YOLOv3

Darknet is an open source neural network framework written in C and CUDA. It is fast, easy to install, and supports CPU and GPU computation.

1. According to the configuration of your computer, edit the Makefile and compile by the command:

cd darknet
make

2. backup file includes all the trained weights:

	action1.weights - 13 pedestrian actions from Stanford Action Dataset - training by YOLOv2
	actionVOC.weights - 8 pedestrian actions from re-built PASCAL VOC Dataset' - training by YOLOv3
	actionVOC_544.weights - improve the resolution and generate the appropriate anchors using k means clustering
	coco_traffic_2017.weights - 14 types of objects in the traffic scenario - collected from COCO Dataset

Please download weights https://drive.google.com/file/d/1vXJTkrWZxue_nqbCpKQK23t48BpTWWxV/view?usp=sharing

3. cfg file includes all the network architectures and the dataset path:

	action1.cfg		action1.data
	actionVOC. cfg		actionVOC.data
	actionVOC_544.cfg	actionVOC_544.data
	coco_traffic_2017.cfg	coco_traffic_2017.data

4. data file includes all the category lists:

	action1.data
	actionVOC. cfg
	actionVOC_544.cfg
	coco_traffic_2017.cfg

5. All the dataset are too large, please download from the following links:

	Modified Stanford Action Dataset: https://drive.google.com/file/d/1w9uiDB1TDKfCcnve7ilYJ1PeXUTYg0OU/view?usp=sharing
	Modified PACSAL VOC Dataset: https://drive.google.com/file/d/1wwztVMlmmwuA8mM5K3460tvnyIizFMbv/view?usp=sharing
	Modified COCO Dataset: https://pan.baidu.com/s/1JKzGP4qGJUUgpn11ywKF2A
	Full COCO Dataset: http://cocodataset.org/#download	- if you want to add or remove some categories you can download 'annotations' 'train2017' 'val2017'

6. Command for detecting using the trained models: (replace '~' with the specific name of the model you want to use)

detect in images:
		./darknet detector test cfg/~.data cfg/~.cfg backup/~.weights test_image.jpg
detect in videos:
		./darknet detector demo cfg/~.data cfg/~.cfg backup/~.weights test_video.mp4

7. Command for training a new model: (replace '~' with the specific name of your model)

by YOLOv3: 
	./darknet detector train cfg/~.data cfg/~.cfg darknet53.conv.74
by YOLOv2:
	./darknet detector train cfg/~.data cfg/~.cfg darknet19_448.conv.23

For more information see the Darknet project website.

About

Object Detection on traffic scenario and Action Recognition (for driverless vehicles) YOLOv3

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C 90.5%
  • Cuda 7.8%
  • Other 1.7%