client

Client

We used RaspPi 4b 16GB RAM with Raspbian 32 bits.

📂 Organization

├── LED.py              # turns on IR ring if connected to GPIO
├── record.py           # record thread
├── watch.py            # processing image thread  
|
├── run.sh              # bash to run capture
├── test.sh             # bash to test the camera stream
├── connect.sh          # start the ptp server
|
├── requirements.txt    # python requirements
|
└── raspividyuv         # modified command to call camera wrapper

🏗️ Requirements

• When installing the Raspbian OS, change the hostname to camX, for X the camera ID number

• Access the Pi remotely (e.g. for camera 0)

ssh [email protected]

• Install requirements

sudo apt-get install ptpd
sudo apt-get install -y libatlas-base-dev libhdf5-dev libhdf5-serial-dev libjasper-dev
pip3 install -r requirements.txt

• Deactivate NTP server

systemctl disable systemd-timesyncd.service

• Start the ptpd server

source connect.sh

• Connect a jumper to the GPIO 4 if you want the RPi to trigger the LED
• Test if the camera stream can be opened

source test.sh

• Check the image using VNC server

• Activate VNC in sudo raspi-config > Interface
• Install VNC viewer in your Desktop and add a new connection with your Rasp IP
• Uncomment hdmi-hot-plug in boot/config.txt if using a Rasp 4
• Activate direct capture under VNC server > Options
• If you receive a warning to unsupported privilege, run sudo vncpasswd -service

• Adjust the blob detectors parameters in watch.py to the size of your blob

1. Change properties of params in watch.py
2. Comment the lines 45 to 54 in record.pyand uncomment line 55
3. Uncomment the lines 61 to 67 in watch.py
4. Run source run.sh to view the captured blob and centroids
5. Repeat steps 1 to 3 until all blobs are detected

⚔️ Usage

1. Run a capture routine in the server (mocaprasp cec, mocaprasp gpe or mocaprasp scr)
2. source run.sh
3. Press Ctrl+C when capture is finished to avoid waiting for the timeout

You can change the timeout waiting in line 85 of watch.py (default: 300 seconds)

🖼️ Example pics