Skip to content

Advanced Object Search

Advanced Object Search is a powerful tool that allows you to filter detections by video, date, object type, object attributes and object movement. It is able to apply two filters at the same time on the same video, to produce results based on complex needs.

Open Queries. Select Advanced Object Search from the list on the left.

Add videos

Click Select Video. Select a video from the list. Click Select date to select a time frame for the analytics to run, depending on the type of video chosen:

  • Earliest: Run the query from the time shown (set during video registration as the beginning timestamp of the video).
  • Latest: Run the query until the time shown (set during video registration as the end timestamp of the video).
  • Given Date: You can select a custom start and end date and time to run the query on. Type the desired date and time or select them by clicking the calendar icon.
  • Live: Analysis is run live, continously.
  • Interval: You can select a custom start and end date and time to run the query on. Type the desired date and time or select them by clicking the calendar icon.
  • Set interval to: Choose from the presets below to run the query on a given time inmterval.
  1. Click Edit object.

Object parameters

IRIS+ Professional is able to filter for 81 object types. They are grouped into five object type categories: Person, Vehicle, Animal, Background and Other.

Select an object type category, then select an object type from the drop-down list. Person is chosen by default.

Click Save to register your Query, or optionally move on to advanced settings.

Some object types may have additional attributes available for filtering (such as helmet and gender for person objects). You can select any number of them to further filter your results; e.g. adding the helmet attribute to the person object type will filter detections so that only people with helmets will show up in the Query results. Multiple attributes can be filtered at the same time (e.g. both female and helmet).

  1. Click the +New Attribute button.
  2. Select an attribute from the list.
  3. Tick the checkbox of the respective attribute to toggle filtering - i.e. whether to look for people with helmets or without them.
  4. You can keep adding additional attributes if the object type allows them - otherwise, click Assign.

If you have a good idea of what pixel range the detected objects would fall into, you can customise the minimum and maximum height of detections, in pixels, here. This would ensure that only objects that fit into that range would get detected.

  1. Click Size.
  2. Adjust the slider's endpoints to set the minimum / maximum values. Note the arrows changing size on the video thumbnail: they can help you gauge the possible height of objects appearing on the video.
  3. Click Assign.

Each objects detected gets assigned a confidence value, ranging from 0% to 100%. Confidence represents the sensitivity of the detector - the minimum and maximum values between which objects are considered to be detected. Increasing this will result in less false positives, but may result in some missed detections, and vice versa.

  1. Click Confidence.
  2. Adjust the slider's endpoints to set the minimum / maximum values.
  3. Click Assign.

Track filters

IRIS+ Professional uses various filters to tell the detector what to look for, such as line crossing, or filter objects by whether they are moving or stationary.

  1. Click the drop-down list of Track filter (Sampling is selected by default).
  2. Select the filter you'd like to use.

Track filters

Takes image samples from the video at the rate defined below, to produce results.

Sampling interval (in milliseconds, 5000 by default): The rate at which samples are taken from the video, to be analysed. Acts as a cooldown parameter.

Produces an event whenever a selected object crosses a line. The line can be one- or bidirectional.

Click the Edit line button Left-click on the camera thumbnail to begin drawing the line; left-click again to finish drawing. Select from the options below the canvas to select the facing direction.

Produces an event whenever an object starts moving after a configurable amount of time spent moving.

Max move pixels (in pixels, 70 by default): The number of pixels that must be in motion for the object to be considered as moving. The higher this value, the less chance for false detections - note, however, that this might also result in some missed detections.

Min stopped duration (in milliseconds, 5000 by default): The minimum amount of time the objects has to stay still to be considered stopped.

Produces an event whenever an object stops moving after a configurable amount of time spent staying still.

Max move pixels (in pixels, 70 by default): The number of pixels that must be in motion for the object to be considered as moving. The higher this value, the less chance for false detections - note, however, that this might also result in some missed detections.

Min stopped duration (in milliseconds, 5000 by default): The minimum amount of time the objects has to stay still to be considered stopped.

Produces an event whenever an object appears on the scene for the first time.

Min track duration (in milliseconds, 2000 by default)": The minimum amount of time the objects must be able to be tracked to be considered appearing.

Filters detections to moving objects only. The minimum movement required to consider an object to be moving is configurable.

Sampling interval (in milliseconds, 5000 by default): The rate at which samples are taken from the video, to be analysed. Acts as a cooldown parameter.

Min move (in percentages, 5 by default): What percentage of pixels, in relation to the object's height, must be in motion for the object to be considered as moving. The higher this value, the less chance for false detections - note, however, that this might also result in some missed detections.

Filters detections to stationary objects only.

Sampling interval (in milliseconds, 5000 by default): The rate at which samples are taken from the video, to be analysed. Acts as a cooldown parameter.

Min move (in percentages, 5 by default): What percentage of pixels, in relation to the object's height, must be in motion for the object to be considered as moving. The higher this value, the less chance for false detections - note, however, that this might also result in some missed detections.

Min stopped duration (in milliseconds, 5000 by default): The minimum amount of time the objects has to stay still to be considered stopped.

Zone drawing (optional)

Zones can be drawn to filter detections based on where they appear in the video. This can be useful for filtering out objects that are not of interest, or for focusing on a specific area of the video.

  1. Click the Edit zone button.
  2. Single-click on the camera thumbnail to draw vertices; double-click to finish drawing.
Advanced settings
  • Name: A name identifying your query. By default, a unique ID is auto-generated, which can be changed here.
  • Description (optional): Describe your Query here.
  • Time aggregation (in milliseconds, 5000 by default): The frequency of sampling; the detector will only search for objects once the time set here has passed after the previous detection.

Adding a second filter

Advanced Object Search is able to apply two filters at the same time on the same video, to produce results based on complex needs. It is possible to mix-and-match two different track filters, object types as well as object parameters. Example use cases:

  • Measure the speed of cars running lights
  • Near miss events
  1. Click the Add second filter button.
  2. Configure your filter to your liking by following the steps above.
  3. Set Join parameters:
  • Max time difference (in seconds, 5 by default): The maximum amount of time between detection of two objects.
  • Max distance (in percentages, 100 by default): The maximum distance between the centres of the two bounding boxes relative to the height of the larger bounding box.
  • Join will be performed on the same track: If selected, the two filters selected will apply to the same object, if applicable. Useful in case you want two different filters applied on the same object - e.g. detecting people who are not wearing protective vests AND are crossing a line.
Example use case configurations

"I would like a Near Miss query which alerts me when a moving car is seen within 4 metres of people, in a 5 second time window."

  • A sampling track filter with person object type.
  • A moving track filter with car object type. To exclude stationary cars.

Join parameters:

  • Max time difference: 5 seconds To look for cars and people visible within 5 seconds of each other.
  • Max distance: 100% To look for cars and people within ~4 metres of each other.
  • Join will be performed on the same track: disabled

"I would like a Query which alerts me when a car is seen not stopping at a pedestrian crossing."

  • A sampling track filter with person object type.
  • A line cross track filter with car object type. A line drawn parallel to the crossing, in both directions. To detect cars not stopping at the crossing.
  • A zone on the person track filter, drawn on the pavement next to the crossing. To filter out people not at the crossings.

Join parameters:

  • Max time difference: 5 seconds To look for cars and people visible within 5 seconds of each other.
  • Max distance: 300% To look for cars and people within a 10-12m distance of each other.
  • Join will be performed on the same track: disabled
  1. Click Save to register your Query.

All Set! Your Query will appear under Queries, in the list on the right.