Skip to content

Releases: glossyio/traffic-monitor

0.6.0 Release

Choose a tag to compare

@glossyio glossyio released this 13 May 05:17
69ff2be

Major changes for 0.6.0

Breaking changes

There are substantial changes to directories in both the repo and system application paths. An update-in-place was not designed and will not work.

The Traffic Monitor software is still undergoing major changes, consider it in Beta and that these major updates will break your current system.

Many configuration defaults have been updated for better performance and new capabilities. It is recommended to start with the new defaults and modify to fit your needs.

Upgrade path

To upgrade, we strongly recommend you backup your settings and database and re-image from scratch. Database structures have not changed, so you should be able to copy over your database from v0.5.

Perform a fresh install by following the Software Installation docs.

Sorry for any inconvenience this may cause.

New Features

Migrated to Podman

We have migrated all containerized applications from Docker to Podman. Podman is a daemon-less architecture that aligns very well with our open ethos while maintaining compatibility with our core applications, such as Frigate.

All previous docker commands are fully available in Podman. Note, though, that since this a daemon-less and runs on systemd, restarts may be handled by the systemctl command.

Thank you for the amazing work on this, @socket-wrench!

Now utilizing H.265 RPi5 hardware video decoding

This was a long-overdue backlog item. The Raspberry Pi 5 contains a hardware decoder for H.265 (h265 / HVEC) videos that allows much more efficient processing of videos encoded with this codec and a smaller size for the same quality as H.264. Be aware that some systems and browsers may not be able to play back H.265 by default.

Frigate sees a substantial decrease in CPU usage: in our tests about 6x (six times) decrease in CPU load. This is particularly obvious in very busy areas with many objects being tracked; i.e. a very busy road with dozens of cars, bikes, and pedestrians. This is achieved by passing the hardware decoder into Frigate and utilizing a custom ffmpeg value in the config.

To utilize the hardware decoder, the Raspberry Pi Camera 3 rtsp feed now encodes to H.265. This uses slightly more CPU than the previous H.264 encoding format since RPi5 has no hardware encoder, but it pays back dividends during Frigate's decoding and storage.

Note that the camera feed for the RPi Cam 3 Wide now properly utilizes the entire field-of-view. Previously the default resolution cropped and stretched the images. This will make the picture substantially clearer and better-looking as well as provide a better view for object detection, especially with far-away / small objects.

All camera feeds, including those from any external cameras, should be specified to send streams to the traffic monitor using H.265/HVEC to take advantage of Raspberry Pi 5's hardware decoder in Frigate.

Dashboard changes

Node-Red Dashboard 2 has updated its internal framework to chart.js, and our charts have been updated to work with the new data format and keep consistency with the on-device dashboards.

Database Download capabilities now include progress bars and allow for very large databases (>> 2 GB) to be properly packaged and transferred in tar.gz format without timing out. JSON files also use the same pipeline.

Thank you for the great, consistent work on this, @mhlmhl!

What's Changed

Full Changelog: v0.5.0...v0.6.0

0.5.0 Release

Choose a tag to compare

@glossyio glossyio released this 11 Mar 17:42
c996826

Major changes for 0.5.0

Breaking Changes

Changes here should not have a major impact on current systems, and there were no database changes. An update-in-place was not designed and may not work.

The Traffic Monitor software is still undergoing major changes, consider it in Beta and that these major updates will break your current system.

Upgrade path

To upgrade, we strongly recommend you backup your settings and database and re-image from scratch. Database structures have changed, so do not restore the databases without first altering the structure--we recommend you start with a blank slate.

Perform a fresh install by following the Software Installation docs.

Sorry for any inconvenience this may cause.

New Features

Automated Detector Installation

Added Raspberry Pi AI HAT+ (Hailo-8 and Hailo-8l) firmware and driver installation! tmsetup will now automatically handle the Coral TPU and the Hailo-8. The script will detect if the AI co-processor is physically installed or if it's already set up and available during setup. The device will need to be restarted in order for installation to complete.

Dashboard 2 AQ and Movement

For those with an Enviro+ Board, the new dashboard will now display sensor values.

Movement patterns are available to validate against manual counts and verify Frigate Zone alignment.

image

Plate Recognizer

The closed-source Plate Recognizer model offers a variety of capabilities that may be of use, including License Plate Recognition, Vehicle Make/Model/Type, vehicle color, and more.

Install by running bash traffic-monitor/script/tmsetup.sh - C plate_recognizer

It will require an internet connection to set up and a paid license with platerecognizer.

What's Changed

Full Changelog: v0.4.0...v0.5.0

0.4.0 Release

Choose a tag to compare

@glossyio glossyio released this 10 Dec 05:00
2fa8d35

Major changes for 0.4.0

Breaking Changes

Major, breaking changes across the board.

The Traffic Monitor software is still undergoing major changes, consider it in Beta and that these major updates will break your current system.

Upgrade path

To upgrade, we strongly recommend you backup your settings and database and re-image from scratch. Database structures have changed, so do not restore the databases without first altering the structure--we recommend you start with a blank slate.

Perform a fresh install by following the Software Installation docs.

Sorry for any inconvenience this may cause.

New Features

New TM Setup Script

We have introduced a new, automated, omnipotent setup process! 🤖

The Traffic Monitor software is installed via an Ansible deploy script executed from our tmsetup shell script. This allows you to perform a local installation or remote installation from a host computer to 1 or more devices simultaneously!

Thank you for @socket-wrench for the amazing work on this!

Node-RED Dashboard 2 introduction

The original Node-RED Dashboard has been deprecated for some time now. We are switching over the on-device dashboard to Dashboard 2. We went this route due to its tight integration with Node-RED, active development, and Vue.js component base. For now, it is accessible via http://<device_ip_address>:1880/dash alongside the previous UI.

Thank you to @mhlmhl for the incredible UI building!

Node-RED moved to Docker

In order to better isolate our primary logic, we have moved it from a system service to another Docker container.

What's Changed

New Contributors

Full Changelog: v0.3.1...v0.4.0

v0.3.1

Choose a tag to compare

@glossyio glossyio released this 02 May 20:29
5c96fe1

What's Changed

No need to upgrade from v0.3.0 if you are already up-and-running.

Full Changelog: v0.3.0...v0.3.1

v0.3.0

Choose a tag to compare

@glossyio glossyio released this 09 Mar 19:28
db2ace1

What's Changed

Major, breaking changes across the board.

To upgrade, we strongly recommend you backup your settings and database and re-image from scratch. Database structures have changed, so do not restore the databases without first altering the structure--we recommend you start with a blank slate.

Highlights

We are still in early-beta with the software and appreciate your patience. This latest release includes the following and more:

  • Increased stability, such as addressing the intermittent WiFi connectivity issue by @socket-wrench, @ernstkl, and @mhlmhl
  • Vast array of system metrics including CPU, RAM, and storage utilization, network interface, power utilization, and more by @mhlmhl. These will help in system monitoring and issue diagnosis. Currently these metrics are only sent to ThingsBoard.
  • Initial config Backup and Restore functionality on ThingsBoard by @mhlmhl and @glossyio. This allows a the device configuration files to be manually backed up to ThingsBoard and restored after, say, a device is re-imaged. This currently requires reconnecting to ThingsBoard after a re-image, and then restore may be triggered via the on-device dashboard UI.
  • Configuration file refactoring to allow for any number an combination of sensors including cameras, radars, air quality monitors, and more by @glossyio. This will also allow for all configuration to be done file-based so it no longer needs to be entered into the on-device dashboard UI. This also introduces Deployment Location information that can be sensor-specific. This will allow, say, if you have multiple cameras to be assigned a unique bear (direction the sensor is facing) on a per-sensor basis. This is now captured in its own database table for better downstream processing.
  • Air Quality telemetry for applicable sensors by @glossyio. Now supports Pimoroni Enviroplus sensor for Temp, Humidity, Barometer, gaseous sensor, particulate matter, and more. The code is stored in a partner repo greendormer/enviroplus-monitor and flow created to capture data locally and to TB backend. Based on the amazing work at roscoe81/enviro-monitor (cloned from 73a8e02, version 8.2 Gen) and the manufactuer's repo at pimoroni/enviroplus-python.
  • Visit our updated documentation: https://docs.trafficmonitor.ai
  • Traffic Monitor Enclosure 3D models are now available at greendormer/tm-enclosure-3d so you can print your own weather-resistant enclosure for the TM!

Thank you everyone for the contributions!

Major changes since v0.2.0

New Contributors

Full Changelog: v0.2.0...v0.3.0

v0.2.0 - IoT hub introduced

Choose a tag to compare

@glossyio glossyio released this 13 Sep 17:38
5610af4

What's Changed

Major, breaking changes across the board to reduce resource utilization and add capabilities.

To upgrade, we strongly recommend you backup databases and completely re-image device to use these features. Database structures have changed, so do not restore the databases without first altering the structure--we recommend you start with a blank slate.

Highlights

We went away from InfluxDB in favor of a SQLite database. This will allow us to avoid the Docker overhead. We found we were not utilizing the Flux language, querying capabilities, and didn't have a use case to need the big-data querying engine.

We went away from Node-RED in a Docker environment and now install Node-RED as a service. The primary motivation here was having system-level access to execute commands from the Node-RED UI, which we intend to be the primary method for users to interact with their device. We wanted to be able to do things like set the system time, create hotspot AP, change WiFi settings, read system-level settings, and more. This was an instance where Docker's isolation system was holding us back.

For our IoT Hub, we went with ThingsBoard after evaluating many options both open-source and closed-source. We are very excited about ThingsBoard's open source nature and ability to provision devices, collect and process data, provide visualizations, and export data for downstream consumption. Anyone can install ThingsBoard by using the Community Edition, using their PaaS Cloud offering, and/or professional edition.

Major changes since v0.1.0

  • Convert InfluxDB to SQLite and facilitate easy database dumps TG#34 by @glossyio in #11
  • Integrated ThingsBoard HTTP API send telemetry by @mhlmhl in #17
  • Added Deployment Location database table and entries by @glossyio in #20
  • Swap Node-RED Docker for Node-RED as a Service TG#89 by @glossyio in #14
  • Added form to change system time settings. by @glossyio in #20
  • Added attrbitues to event payload by @glossyio in #20

Minor changes since v0.1.0

  • Bumped Frigate to v0.14.0 by @glossyio in #20
  • Modified default Frigate Bicycle Object Filter, drastically lowered score to have less false negatives by @glossyio in #20
  • added example nmcli commands for connect to wifi (disable AP mode) by @glossyio in #2
  • add apt update for coral tpu and change to full-upgrade by @glossyio in #3
  • Added description, better starter steps by @glossyio in #5
  • fixed envsubst not adding HOSTNAME to frigate config by @glossyio in #7
  • Feature/pci tpu and telemetry by @glossyio in #8
  • Feature/nr db download links by @glossyio in #9
  • fixed node-red install command by @glossyio in #15
  • Kr 301 27 config tab by @mhlmhl in #17

New Contributors

Full Changelog: v.0.1.0...v0.2.0

v0.2.0-beta.3

v0.2.0-beta.3 Pre-release
Pre-release

Choose a tag to compare

@glossyio glossyio released this 16 Aug 13:52
8a49971

What's Changed

Full Changelog: v0.2.0-beta.2...v0.2.0-beta.3

v0.2.0-beta.2

v0.2.0-beta.2 Pre-release
Pre-release

Choose a tag to compare

@glossyio glossyio released this 07 Aug 18:27
afb009b

What's Changed

  • Swap Node-RED Docker for Node-RED as a Service TG#89 by @glossyio in #14

Full Changelog: v.0.2.0-beta.1...v0.2.0-beta.2

v.0.2.0-beta.1

v.0.2.0-beta.1 Pre-release
Pre-release

Choose a tag to compare

@glossyio glossyio released this 23 Jul 17:40
8e9b0a4

What's Changed

  • added example nmcli commands for connect to wifi (disable AP mode) by @glossyio in #2
  • add apt update for coral tpu and change to full-upgrade by @glossyio in #3
  • Added description, better starter steps by @glossyio in #5
  • fixed envsubst not adding HOSTNAME to frigate config by @glossyio in #7
  • Feature/pci tpu and telemetry by @glossyio in #8
  • Feature/nr db download links by @glossyio in #9
  • Convert InfluxDB to SQLite and facilitate easy database dumps TG#34 by @glossyio in #11

New Contributors

Full Changelog: v.0.1.0...v.0.2.0-beta.1

v.0.1.0 - initial release

Choose a tag to compare

@glossyio glossyio released this 26 Jun 14:58
3dcaba8