-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.md.bak
97 lines (51 loc) · 4.28 KB
/
README.md.bak
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
# RobotSorting system using Kinect (Under-developed)
This project implements a device for sorting common subjects using robot and Kinect.
We build our recognition method on YOLO_v3, for more infomation on YOLO, see https://github.com/pjreddie/darknet.
1. HOW TO INSTALL
2. HOW TO USE
- To get started, you should make sure you have already installed the following dependancies
- OpenCv
- CUDA 9.0
- CUDNN for Cuda 9.0
- Microsoft Visual Studio
- And all the file at this repo (RobotSorting)
I will not follow you through all this and it WILL BE PAINFUL.
Note that if you are using Linix, this project is NOT suitable for you.
- After downloading all the dependancies mentioned above, you can use VS2015 ( Not tested on other versions ) to make your
darknet.exe file at "darknetpro\darknet\build\darknet\x64".
In CMD, after entering "darknetpro\darknet\build\darknet\x64", use the following command:
/* darknet.exe detector test data/coco.data yolov3.cfg yolov3.weights -i 0 -thresh 0.25 */
If you see something like this:
It means you successfully downloaded YOLO3 on your platform.
- The program will search in file
darknetpro\darknet\build\darknet\x64\picpath
to get your test picture, make sure you have named them like test_1 test_2 and so on.
These file contain the address of your test pictures. For those pictures, put them in:
D:\git_project\yolo\darknetpro\darknet\build\darknet\x64\testpic
For a test, we used this picture:
After doing so, run the darknet again and you will get output like this:
And a picture after prediction at
D:\git_project\yolo\darknetpro\darknet\build\darknet\x64
Like this:
/* Note that there are several class we won't use in our future work, for example the DININGTABLE in the picture */
/* We use COCO to train our model which contains more than 80 classes of object, here I list all of them below */
[¡®person¡¯, ¡®bicycle¡¯, ¡®car¡¯, ¡®motorcycle¡¯, ¡®airplane¡¯, ¡®bus¡¯, ¡®train¡¯, ¡®truck¡¯, ¡®boat¡¯, ¡®traffic light¡¯, ¡®fire hydrant¡¯, ¡®stop sign¡¯, ¡®parking meter¡¯, ¡®bench¡¯, ¡®bird¡¯, ¡®cat¡¯, ¡®dog¡¯, ¡®horse¡¯, ¡®sheep¡¯, ¡®cow¡¯, ¡®elephant¡¯, ¡®bear¡¯, ¡®zebra¡¯, ¡®giraffe¡¯, ¡®backpack¡¯, ¡®umbrella¡¯, ¡®handbag¡¯, ¡®tie¡¯, ¡®suitcase¡¯, ¡®frisbee¡¯, ¡®skis¡¯, ¡®snowboard¡¯, ¡®sports ball¡¯, ¡®kite¡¯, ¡®baseball bat¡¯, ¡®baseball glove¡¯, ¡®skateboard¡¯, ¡®surfboard¡¯, ¡®tennis racket¡¯, ¡®bottle¡¯, ¡®wine glass¡¯, ¡®cup¡¯, ¡®fork¡¯, ¡®knife¡¯, ¡®spoon¡¯, ¡®bowl¡¯, ¡®banana¡¯, ¡®apple¡¯, ¡®sandwich¡¯, ¡®orange¡¯, ¡®broccoli¡¯, ¡®carrot¡¯, ¡®hot dog¡¯, ¡®pizza¡¯, ¡®donut¡¯, ¡®cake¡¯, ¡®chair¡¯, ¡®couch¡¯, ¡®potted plant¡¯, ¡®bed¡¯, ¡®dining table¡¯, ¡®toilet¡¯, ¡®tv¡¯, ¡®laptop¡¯, ¡®mouse¡¯, ¡®remote¡¯, ¡®keyboard¡¯, ¡®cell phone¡¯, ¡®microwave¡¯, ¡®oven¡¯, ¡®toaster¡¯, ¡®sink¡¯, ¡®refrigerator¡¯, ¡®book¡¯, ¡®clock¡¯, ¡®vase¡¯, ¡®scissors¡¯, ¡®teddy bear¡¯, ¡®hair drier¡¯, ¡®toothbrush¡¯]
Back to the point, you will get an output file which contains the pixel position of the Bounding Box (we take the x and y of the center
pixel for future work) at:
\darknetpro\darknet\build\darknet\x64\box_xy
Like this:
- Connect Kinect to YOLO
1. Download TestKincet in this project, all of the source files are in it.
2. Connect the Kincet to your device and run the code at \TestKinect.
3. The program will take photos and save them at
\darknetpro\darknet\build\darknet\x64\testpic
4. Now start YOLO_V3, it will search images in this file and make preditions.
5. Program at TestKinect would search the outputfile at \darknetpro\darknet\build\darknet\x64\box_xy to get the position of
the box and calculate its correspoding depth (the distance of the objects)
Eventually you have this:
