Repository
https://github.com/shirokunet/RCVehiclePy
Videos
Category: Robotics
-
Run WebRTC Native Client Momo on Raspberry Pi 3B
Goal
Download momo binary and run on Raspberry Pi 3B to see a Pi’s camera image on Host PC.
Environment
- Server: Raspberry Pi 3B
- Client: X220 (Ubuntu 18.04)
- Momo: 19.07.0-rc0
Setup Camera
- Connect a camera to Pi
- $ vcgencmd get_camera
- supported=1 detected=1 # Should be like this
Setup Library
- $ sudo apt-get install libnspr4 libnss3
- $ sudo modprobe bcm2835-v4l2
- $ sudo vi /etc/modules
- bcm2835-v4l2 # For Using Raspberry Pi Camera
- $ sudo reboot
Check IP address
- $ ifconfig
- wlan0: flags=4163
- inet 10.0.0.88 netmask 255.255.255.0 broadcast 10.0.0.255
- wlan0: flags=4163
Get momo
- $ wget https://github.com/shiguredo/momo/releases/download/19.07.0-rc0/momo-19.07.0-rc0_raspbian-stretch_armv7.tar.gz
- You can see newest version at here.
- Raspbian Stretch ARMv7
- For Raspberry Pi 3 B/B+
- Raspbian Stretch ARMv6
- For Raspberry Pi Zero W/WH
- Raspbian Stretch ARMv7
- You can see newest version at here.
- $ tar xvzf momo-19.07.0-rc0_raspbian-stretch_armv7.tar.gz
- $ sudo rm -r momo-19.07.0-rc0_raspbian-stretch_armv7.tar.gz
Run momo
- $ cd momo-19.07.0-rc0_raspbian-stretch_armv7/
- $ ./momo --version
- WebRTC Native Client Momo version 19.07.0-rc0 USE_MMAL_ENCODER=1
- $ ./momo --help
- $ ./momo --no-audio --resolution HD --framerate 30 p2p
Acces
- Go to http://10.0.0.88:8080/html/p2p.html on Host PC
- Select H264
- And push Connect
Benchｍark
The delay is about 100~300msec.
momo 19.02.0
momo 19.07.0-rc0
My Note
- According to the wiki,
- RPi 3 Model B is ARMv8-A (64/32-bit).
- RPi zero is ARMv6Z (32-bit)
- However if we run the command on Pi 3B,
- $ lscpu
- Architecture: armv7l
- Byte Order: Little Endian
- CPU(s): 4
- On-line CPU(s) list: 0-3
- Thread(s) per core: 1
- Core(s) per socket: 4
- Socket(s): 1
- Model: 4
- Model name: ARMv7 Processor rev 4 (v7l)
- CPU max MHz: 1200.0000
- CPU min MHz: 600.0000
- BogoMIPS: 38.40
- Flags: half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm crc32
- $ lscpu
- Which mean Pi 3B is not compatible ARMv8 yet.
Reference
- https://github.com/shiguredo/momo
- https://github.com/shiguredo/momo/blob/develop/doc/USE.md
- https://gist.github.com/voluntas/51c67d0d8ce7af9f24655cee4d7dd253
- https://en.wikipedia.org/wiki/Raspberry_Pi
- https://www.youtube.com/watch?v=9cQT4urTlXM
-
Raspberry Pi 3B Setup
Goal
Install Lite Raspbian on Raspberry Pi 3B, and connect to Pi from Host PC using SSH.
Environment
- Host PC: Ubuntu 16.04
- Client PC: Raspberry Pi 3B
Prepare SD card using Host PC
- Get Etcher to burn the image file.
- Get the Raspbian Stretch Lite from raspberrypi.org.
- Burn zip image file to a blank SD card.
Boot Raspberry Pi
- Insert the SD card your prepared on last step.
- Make sure RasPi should be connected a HDMI display, a keyboard and a mouse.
- Connect a usb cable to RasPi for providing power.
Login
- User: pi
- Password: raspberry
Setup Config (Optional)
- $ sudo raspi-config
- Advanced Options -> Expand Filesystem -> OK
- Localisation Options -> Change Timezone -> Select your location
- Localisation Options -> Change Keyboard Layout -> Generic 101-key PC -> Other -> English(US) -> English(US) -> The default for the keyboard layout -> No compose key
- Interfacing Options -> P2 SSH -> Yes -> OK
- Interfacing Options -> P1 Camera -> Yes -> OK
- Interfacing Options -> P5 I2C -> Yes -> OK
- Finish -> Yes
- $ sudo reboot now
Update Raspbian Package (Optional)
- $ sudo apt-get update
- $ sudo apt-get upgrade
- $ sudo apt-get dist-upgrade
- $ sudo apt-get install -y rpi-update
- $ sudo rpi-update
- $ sudo reboot now
Install Tools (Optional)
- $ sudo apt-get install vim
- $ sudo apt-get install git
Setup Wifi (Optional)
- $ sudo vim /etc/wpa_supplicant/wpa_supplicant.conf
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev update_config=1 network={ ssid="xxxx" psk="xxxx" }Connect to Raspberry Pi from Host PC via Wifi using SSH
- $ sudo ssh pi@raspberrypi.local
Finish!
Reference
-
Robust Highway Lane Segmentation Based on LaneNet Trained BDD100K
This is my final report of TensorFlow class at UCSC.
Abstract
In this report, we describe that how to prepare dataset and improve the performance of the CNN based lane segmentation by using transfer learning. In addition, we evaluate and compare four types of the methods for lanes segmentation such as the simple rule based method, the CNN based method trained with daytime dataset, the CNN trained daytime dataset filtered gamma correction, and the CNN trained high diversity road dataset.
(more…)
