OpenMV Development Guide

Develop wireless sensing applications with OpenMV cameras using Bluetooth Low Energy and Pareto Anywhere.

Prerequisites

An OpenMV camera & IDE and, optionally, a Pareto Anywhere instance to make sense of the data

• 

• Buy a camera from OpenMV

  Support the ongoing open source development of the OpenMV ecosystem!

• 

• Download the OpenMV IDE

  Install the integrated development environment for OpenMV cameras.

• 

• [OPTIONAL] Run Pareto Anywhere on a PC

  Our step-by-step guide to install and run on a Windows/Mac/Linux PC (or server).

Connecting to the OpenMV camera   Step 1 of 3

Connect with the camera in the OpenMV IDE.

---

How to connect?

A USB connection from your PC powers the camera and enables serial communication with the IDE.

What's an IDE?

An Integrated Development Environment facilitates the development and testing of code, in this case directly on the OpenMV camera itself.

Prepare the OpenMV camera Part 1

  Your OpenMV camera will likely ship ready to connect to the OpenMV IDE.

If in doubt, consult the official documentation to ensure that any firmware updates are installed, and that the hardware (jumpers, antennas, ...) is correctly configured.

Connect via USB cable Part 2

Connect in the OpenMV IDE Part 3

Run the OpenMV IDE on your PC. The default view should resemble the following. Annotations have been added to highlight key features.

Click the Connect icon to connect to the camera via USB. The OpenMV IDE footer will display information about the device which is currently connected, in this case the OpenMV AE3 with the PAG7936 sensor.

The camera can now be programmed via the OpenMV IDE, which is the subject of the next step.

Programming the OpenMV camera   Step 2 of 3

Run a Python script on the camera.

---

Why Python?

OpenMV uses the MicroPython operating system to run Python 3 scripts, which make working with machine vision algorithms much easier.

Any documentation?

Yes, consult the MicroPython for OpenMV documentation to learn more.

Run helloworld.py Part 1

The OpenMV IDE should open a helloworld.py file by default. Click the Start (run script) icon at the bottom left of the IDE to run this script on the camera.

Observe the sensor snapshots updating live in the Frame Buffer and the frames per second calculations printed in the Serial Terminal.

  Click the Stop (halt script) icon at the bottom left of the IDE to stop a running script.

Run OpenMV examples Part 2

Explore the extensive capabilities of the OpenMV camera by running the example scripts provided in the IDE. For instance, to test out the machine learning capabilities for face detection using TensorFlow, select the following:

File ➙ Examples ➙ Machine Learning ➙ TensorFlow ➙ blazeface_detector.py

  Some examples are camera-specific and might not run on yours!

Transmitting data   Step 3 of 3

Relay machine vision and sensor data wirelessly.

---

Using BLE?

Yes, Bluetooth Low Energy is ideal for battery-powered applications that require periodic transmission of a few bytes of data (ex: sensor readings).

Using WiFi?

Yes, WiFi facilitates the transmission of larger data (ex: images), with the tradeoff of higher energy consumption.

Transmit BLE advertising packets Part 1

The MicroPython bluetooth module includes a BLE class which provides a gap_advertise function to specify and transmit Bluetooth Low Energy advertising packets. The two key parameters of the gap_advertise function are:

• interval_us   Interval, in microseconds, between advertising packet transmissions.
• adv_data   The GAP^* data, to include in the advertising packet.

^* The Bluetooth Generic Access Profile (GAP) defines the advertising packet data structure.

The following Python script will advertise, every 0.5 seconds, OpenMV as the local name of the device.

import bluetooth, time

_ADV_PAYLOAD = [ 0x07, 0x09, ord('O'), ord('p'), ord('e'), ord('n'), ord('M'), ord('V') ]

ble = bluetooth.BLE()
ble.active(True)
ble.gap_advertise(500000, adv_data=bytearray(_ADV_PAYLOAD))

while(True):
    time.clock().tick()

Copy Ctrl-C and paste Ctrl-V the code into a new file in the OpenMV IDE and then click the Start (run script) icon.

Observe in Pareto Anywhere, or from any Bluetooth Low Energy scanning mobile app, a device named OpenMV. This is the camera advertising its name to any devices in range.

Replace the _ADV_PAYLOAD from the script above with any of the examples below, in order to transmit different types of data.

• +

# Battery Level
# Advertising payload bytes:
# 0:   0x04 = Length (not including the length byte itself)
# 1:   0x16 = 16-bit service data (from GAP)
# 2-3: 0x2a19 = Battery Level service UUID (from GATT)*
# 4:   0x64 = Battery level as percentage (0x64 = 100%)
batteryPercentage = 100
_ADV_PAYLOAD = [ 0x04, 0x16, 0x19, 0x2a, batteryPercentage ]

# Generic Level (ex: gauge reader)
# Advertising payload bytes:
# 0:   0x05 = Length (not including the length byte itself)
# 1:   0x16 = 16-bit service data (from GAP)
# 2-3: 0x2af9 = Generic Level service UUID (from GATT)*
# 4-5: 0x1234 = Generic level as unitless 16-bit value*
levelMsb = 0x12
levelLsb = 0x34
_ADV_PAYLOAD = [ 0x05, 0x16, 0xf9, 0x2a, levelLsb, levelMsb ]

# Distance (ex: ToF sensor)
# Advertising payload bytes:
# 0:   0x07 = Length (not including the length byte itself)
# 1:   0x16 = 16-bit service data (from GAP)
# 2-3: 0xfcd2 = BTHome service UUID*
# 4:   0x40 = BTHome device information
# 5:   0x40 = Distance (mm) Object ID
# 6-7: 0x04d2 = Distance measurement (1234mm)*
distanceMsb = 0x04
distanceLsb = 0xd2
_ADV_PAYLOAD = [ 0x07, 0x16, 0xd2, 0xfc, 0x40, 0x40, distanceLsb, distanceMsb ]

# Motion + Occupancy
# Advertising payload bytes:
# 0:   0x08 = Length (not including the length byte itself)
# 1:   0x16 = 16-bit service data (from GAP)
# 2-3: 0xfcd2 = BTHome service UUID*
# 4:   0x40 = BTHome device information
# 5:   0x21 = Motion Object ID
# 6:   0x01 = Motion detected (1 = true)
# 7:   0x23 = Occupancy Object ID
# 8:   0x00 = Occupancy not detected (0 = false)
isMotionDetected = 1
isOccupancyDetected = 0
_ADV_PAYLOAD = [ 0x08, 0x16, 0xd2, 0xfc, 0x40, 0x21, isMotionDetected, 0x23, isOccupancyDetected ]

Consult our best practices guide for more details and examples:

• 

• Best practices for BLE advertising data

  Structuring Bluetooth Low Energy (BLE) advertising data for interoperability and interpretability.

Transmit images over WiFi Part 2

Details to come using the record-silo of Pareto Anywhere.

Tutorial prepared with ♥ by jeffyactive.

You can reelyActive's open source efforts directly by contributing code & docs, collectively by sharing across your network, and commercially through our packages. We invite you to become a sponsor to OpenMV to their work too!

Where to next?

Continue exploring our open architecture and all its applications.

• 

• Directory of Devices

  Browse all device configuration tutorials and development guides.

• 

• reelyActive Developers

  Browse all developer documentation and tutorials.

• 

• reelyActive

  Together, let's make sense of things.