Welcome to the guide on how to analyse offline data and start collecting online data from live streams using the GoodVision Video Insights platform. If you would like to learn more about what precedes it and what the requirements for connecting GoodVision Live traffic are, read our “Requirements for GoodVision Live Traffic” article or contact us at firstname.lastname@example.org.
Offline data analysis
Data collection for Traffic video streams begins as soon as the video stream is assigned to the data source, as was mentioned in the configuration guide The offline data analysis process for Traffic video streams is the same as for cloud-processed videos.
The Describe page provides the ability to specify which traffic flows you wish to analyze by drawing virtual lines and zones to indicate entries, exits, and other areas of interest within your traffic scene. In addition, you can define various traffic movements, events, and complex scenarios as combinations of these lines and zones.
After describing the scene for analysis, you can create analytical widgets to display aggregated outputs of your movements and events. These widgets provide predefined metrics and visual maps, such as traffic counts on movements, peak hours, average travel times, time gaps, and various heatmaps. For more information on creating analytical widgets, please refer to our guide.
When working with ANPR video stream, you will not see a standard data source set up, to which you are used to. In this case, you will work only with live data that will start collecting after you define traffic scenarios. Further information about live data collection can be found in the following text.
Live data collection and analysis
Before you can begin collecting live data, it is necessary to define your traffic scenarios in the Live Dashboard section. Traffic scenarios refer to specific traffic events that you want to monitor on the road. GoodVision's platform offers a variety of statistical events and violations that you can monitor, including cross directions, average speed, illegal stops, and licence plate detection, with more options set to be added in the future.
Depending on the selected scenario type, you will need to define a few scenario parameters. These may include defining areas that objects should cross or remain in, selecting object classes that should be detected, setting the distance for speed calculation, etc.
As the final step in scenario configuration, you can create as many analytical widgets as needed to provide visual representations of the data collected during traffic monitoring. This will enable you to gain valuable insights and make informed decisions based on the data.
To create a new widget, simply click on the NEW WIDGET button and select the visualisation you need. A preview of the created widget will appear on the page with default parameters such as widget name, time range, and unit. You can easily change them with just a click if needed.
It's important to note that once a traffic scenario is defined we start collecting data exactly for the defined scenario and not collecting any other events. This is the reason why the scenario cannot be changed afterwards but you can only duplicate it and start collecting a new set of data. It's essential to double-check all scenario parameters before saving. Only the scenario name and colour will remain editable in the Traffic Scenario Detail section. Newly created traffic scenario will appear in the right panel.
If you forgot to create all the necessary widgets for your scenario, don't worry. You can add them at any time by using the Create new widget button located in the top left corner of the live dashboard. Additionally, you can create widgets directly from the scenario bean using the Create widgets button or from the Scenario detail page, which provides a better overview of the widgets you have already created. On the dashboard, you can quickly identify which widget belongs to which scenario by looking for the colourful stripe on the left side of the widget.
Live data alerts and integration
If you need to get alerts about detected events in third-party applications, you can create a Live Traffic integration on the Analyze Data page. You can choose whether to receive instant or periodic alerts if alert messages should include images, which object classes should be detected, and set the link where data will be sent.