pinto0309/tinyyolo

Challenge the marginal performance of YoloV2 + Neural Compute Stick + RaspberryPi　YoloV2+Neural Compute Stick(NCS)+Raspberry Piの限界性能に挑戦

[Japanese] TinyYolo

YoloV2+Neural Compute Stick(NCS)+Raspberry Piの限界性能に挑戦　Challenge the marginal performance of YoloV2 + Neural Compute Stick + Raspberry Pi

https://qiita.com/PINTO/items/db3ab44a3e2bcd87f2d8

動作イメージ

TinyYolo + Neural Compute Stick + RaspberryPi3

Youtube: https://youtu.be/L4RvVLyo8Rc

環境

・RaspberryPi 3 + Raspbian Stretch

・NCSDK v1.12.00

・Intel Movidius Neural Compute Stick　１本

・OpenCV 3.4.1

・OpenGL

・numpy

・UVC対応のUSB-Webカメラ

環境構築

1. SWAP領域の一時的な拡張

$ sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=2048

$ sudo /etc/init.d/dphys-swapfile restart swapon -s

2. パッケージのインストール

$ sudo apt-get update
$ sudo apt-get upgrade
$ sudo apt-get install python3-pip python3-numpy git cmake

3. NCSDKのインストール

$ cd ~
$ git clone https://github.com/movidius/ncsdk.git
$ cd ncsdk
$ make install

4. OpenCVのインストール

$ wget https://github.com/PINTO0309/OpenCVonARMv7/blob/master/libopencv3_3.4.1-20180304.1_armhf.deb
$ sudo apt install -y ./libopencv3_3.4.1-20180304.1_armhf.deb
$ sudo ldconfig

5. OpenGLのインストール

$ sudo apt-get install python-opengl
$ sudo -H pip3 install pyopengl
$ sudo -H pip3 install pyopengl_accelerate
$ sudo raspi-config

6. 「7.Advanced Options」-「A7 GL Driver」-「G2 GL (Fake KMS)」の順に選択し、Raspberry Pi のOpenGL Driver を有効化

7. 再起動

$ sudo reboot

8. リソース一式のダウンロード

$ cd ~
$ git clone https://github.com/PINTO0309/TinyYolo.git

9. USB-WEBカメラ(UVC対応) と Neural Compute Stick をRaspberryPiのUSBポートへ接続(Neural Compute Stickをマルチで使用する場合は電圧が不足するためセルフパワーUSB-Hub必須)

10. RaspberryPiとディスプレイをHDMIケーブルで接続

11. MultiStick.pyの実行

$ cd TinyYolo
$ python3 ./detectionExample/MultiStick.py

12. SWAP領域の縮小

$ sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=100

$ sudo /etc/init.d/dphys-swapfile restart swapon -s

　 　

[English] TinyYolo

Challenge the marginal performance of YoloV2 + Neural Compute Stick + Raspberry Pi

https://qiita.com/PINTO/items/db3ab44a3e2bcd87f2d8

Image of motion

TinyYolo + Neural Compute Stick + RaspberryPi3

Youtube: https://youtu.be/L4RvVLyo8Rc

Environment

・RaspberryPi 3 + Raspbian Stretch

・NCSDK v1.12.00

・Intel Movidius Neural Compute Stick　１本

・OpenCV 3.4.1

・OpenGL

・numpy

・(UVC)USB-Web Camera

Building environment

1. Temporary extension of SWAP area

$ sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=2048

$ sudo /etc/init.d/dphys-swapfile restart swapon -s

2. Installing packages

$ sudo apt-get update
$ sudo apt-get upgrade
$ sudo apt-get install python3-pip python3-numpy git cmake

3. Installing NCSDK

$ cd ~
$ git clone https://github.com/movidius/ncsdk.git
$ cd ncsdk
$ make install

4. Installation of OpenCV

$ wget https://github.com/PINTO0309/OpenCVonARMv7/blob/master/libopencv3_3.4.1-20180304.1_armhf.deb
$ sudo apt install -y ./libopencv3_3.4.1-20180304.1_armhf.deb
$ sudo ldconfig

5. Installing OpenGL

$ sudo apt-get install python-opengl
$ sudo -H pip3 install pyopengl
$ sudo -H pip3 install pyopengl_accelerate
$ sudo raspi-config

6. 「7.Advanced Options」-「A7 GL Driver」-「G2 GL (Fake KMS)」 and Activate Raspberry Pi's OpenGL Driver

7. Reboot

$ sudo reboot

8. Download complete set of resources

$ cd ~
$ git clone https://github.com/PINTO0309/TinyYolo.git

9. Connect USB-WEB camera (UVC compatible) and Neural Compute Stick to RaspberryPi's USB port (self power USB-Hub required due to insufficient voltage when using Neural Compute Stick in multiple)

10. Connect RaspberryPi and display with HDMI cable

11. Running MultiStick.py

$ cd TinyYolo
$ python3 ./detectionExample/MultiStick.py

12. Reducing the SWAP area

$ sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=100

$ sudo /etc/init.d/dphys-swapfile restart swapon -s

tinyyolo's Issues

Mesa loader issue.

Whenever I ran multistick.py script I got following error even after reinstalling libdrm using sudo apt install libdrm-*. I got live streaming without detection. Pls help me to solve this.

pi@Zero:~/TinyYolo $ python3 ./detectionExample/MultiStick.py

Device 0 Address: 1.2 - VID/PID 03e7:2150
1
Starting wait for connect with 2000ms timeout
Found Address: 1.2 - VID/PID 03e7:2150
Found EP 0x81 : max packet size is 512 bytes
Found EP 0x01 : max packet size is 512 bytes
Found and opened device
Performing bulk write of 865724 bytes...
Successfully sent 865724 bytes of data in 95.934476 ms (8.606069 MB/s)
Boot successful, device address 1.2
Found Address: 1.2 - VID/PID 03e7:f63b
done
Booted 1.2 -> VSC

Loaded Graphs!!!
libGL error: MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
press 'q' to quit!

RuntimeError: module compiled against API version 0xb but this version of numpy is 0xa
/usr/local/lib/python3.5/dist-packages/OpenGL/arrays/numpymodule.py:30: DeprecationWarning: The 'warn' method is deprecated, use 'warning' instead
"Unable to load numpy_formathandler accelerator from OpenGL_accelerate"
Unable to load numpy_formathandler accelerator from OpenGL_accelerate

Finished

Thanks in advance
Hashir

Camera problem

Hi @PINTO0309 , thanks for the awesome repository. It works fine with ip camera , raspberry pi camera v2 but when it comes to usb camera(i have tried two different cameras from microsoft and logitecth) it does not detect anything but i can still see the live stream. To be more specific it goes through the following :

global lastresults

    s, img = cam.read()
    #out.write(img)
    
    print(s)
    print("1")
    if not s:
        print("Could not get frame")
        return 0

    lock.acquire()
    if len(frameBuffer)>10:
        for i in range(10):
            del frameBuffer[0]
    frameBuffer.append(img)
    lock.release()
    res = None
    print("2")

then jumps to :

else:
        if lastresults == None:
            img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
            h, w = img.shape[:2]
            glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, w, h, 0, GL_RGB, GL_UNSIGNED_BYTE, img)

which means lastresults is none and res is none.
So what should i do? , How can i solve this problem.
Looking for your reply