Operating mode «Super Event»#
The operating mode «Super Event» is available in CARS Analytics, which is necessary to create events of detection and recognition of vehicle and license plate attributes with aggregated data from several video streams.
Please note, after enabling the «Super Event» mode, the classic mode of operation of LUNA CARS Analytics will not be available!
Enabling the operating mode «Super Event»#
After the installation of LUNA CARS (see documentation LUNA CARS Installer Installation Manual), the operating mode «Super Event» is deactivated by default. In order to work with it, it is necessary to activate it, for this purpose:
1․ Open the working directory:
cd cars-installer_v.*.*.*/
2․ Make sure that Docker containers are turned off:
docker ps
The list should not contain containers of the LUNA CARS system
3․ If there are LUNA CARS containers in the list, run a script in the working directory to stop them:
./docker_stop_all.sh
Wait until all Docker containers are stopped
4․ Open the «.env» configuration file in the editor:
nano .env
5․ Set the following values for the variables:
ANALYTICS_AGGREGAROR_CMD=start_checkpoint_aggregator
ANALYTICS_PLUGIN_NAME=checkpoint_plugin
ANALYTICS_CHECKPOINT_MODE=True
6․ Save and close the «.env» file
7․ Execute the command to clear the migration file:
rm /var/lib/luna/cars/docker/volumes/media/db_check/*-v.X.X.XX-vX.X.XX
Where v.X.X.XX.XX-vX.X.XX.XX is part of the value of the ANALYTICS_BACKEND_VERSION variable specified in the «.env» configuration file:
ANALYTICS_BACKEND_VERSION=42302e6-v.3.1.21-v1.0.27
8․ Start Docker containers:
./docker_start_all.sh
9․ Open the LUNA CARS Analytics web interface at the following address:
http://ip_address:8080
The port is configured in the
ANALYTICS_FRONTEND_PORTvariable in the «.env» configuration file, the default value is "8080"
Disabling the operating mode «Super Event»#
At the moment, there is no possibility to disable the «Super Event» mode. To operate LUNA CARS Analytics in classic mode, a reinstallation of the system is required (see documentation LUNA CARS Installer Installation Manual).
Available handlers#
Handlers must be defined for generating detection and recognition events, see «Scenarios» section
The operating mode «Super Event» consists of one mandatory handler(Front tracking) and one (or more) additional handlers.
The following handlers are available in operating mode «Super Event»:
Table 42. Description of handlers as part of «Super Event»
| Handler | Purpose | Description | Camera |
|---|---|---|---|
| Front tracking | Forms front view of the vehicle, recognizes license plate from the front | Required handler. Allowed 1 instance | RTSP stream |
| Back tracking | Forms rear view of vehicle, recognizes rear license plate | Optional handler. Multiple instances are allowed | RTSP stream |
| Back crawler | Forms rear view of the vehicle, necessary for refining the parameters of recognition from other angles | Optional handler. Multiple instances are allowed | ANPR camera |
| Side crawler | Forms the side view of the vehicle, defines the number of axes of the vehicle | Optional handler. Multiple instances are allowed | ANPR camera |
«Front tracking» handler#
The parameters of the handler are given in Table 43.
Table 43. «Front tracking» handler parameters
| Name | Description | Possible values | Default |
|---|---|---|---|
| Camera | Video stream frame source | Cameras added to the system | - |
| Region for intersection | A drop-down list for selecting the zone with which the vehicle intersections are checked for the scenario to work. | Created zones for the selected camera. The process of creating recognition zones is described in section Recognition zone creation. | - |
| Min threshold intersection | The ratio of the zone of the BBox of the vehicle or the LP that crossed the Region for intersection to the non-crossed one. It is not recommended to enter more than three decimal places. | 0,01…1,00 | - |
| - 0,01 – the event appears immediately when BBox appears in the Region for intersection | |||
| - 1,00 – the event appears only when BBox is completely overlapped with Region for intersection | |||
| Count of bestshots | The minimum number of frames of the current track to select the best frame. When the minimum is reached, the probability of correct recognition of all attributes and symbols of the LP is higher. If there are fewer frames in the track than the minimum, then the best frame will be selected, but the probability of correct recognition of all attributes and symbols of the LP is lower. The more frames are collected to select the best, the more accurate the recognition of vehicle attributes will be. An increase in the number of frames can lead to a decrease in performance. | 1…1000000 | 1 |
| - 1 – high probability of getting incorrect recognition results | |||
| - 1000000 – the time to select the best frame increases, but the quality is higher | |||
| Max count of bestshots | The maximum number of best frames for one track. When the maximum is reached, the selection of the best frames is completed. | 1…1000000 | 43200 |
| - 1 – high probability of getting incorrect recognition results | |||
| - 1000000 – the time to select the best frame increases, but the quality is higher | |||
| Max delay first bestshot | Pause between the start of the track and sending the first frame to the CARS API to determine the best frame. It is necessary to increase the pause if the vehicle in the first seconds of the track may not be in the optimal angle. | 0…1000 | 0 |
| - 0 – no delay | |||
| - 1…1000 – pause in ms | |||
| Bestshot send periodicity | Frequency of sending bestshots | 1000…1000000 | 2000 |
| - 0 – no delay | |||
| - 1…1000000 – pause in ms | |||
| First control area | Detection zone, by using which the direction of movement of the vehicle is determined | Zones added to the camera | - |
| Second control zone | Detection zone, by using which the direction of movement of the vehicle is determined | Zones added to the camera | - |
| Use track area | Using an additional Detection Zone to work with a part of the road (traffic lane, entrance/exit, stop, etc.). Forming a list of the best frames will only be available for vehicles that overlap the traffic area with a specified threshold. | On/Off | Off |
| Track area | The detection zone with which the vehicle intersections are checked. The introduction of an additional zone will allow you to cut off events that do not need to be recorded. | Created zones for the selected camera. The process of creating zones is described in section Recognition zone creation. | - |
| Track area threshold | The relative area of the Recognition Zone overlapped by the BBox of the vehicle, when exceeding the threshold of which the vehicle will be taken into account to form the best frames. The higher the threshold, the more accurate the detection of the vehicle and LP will be, but the number of passes will also increase. It is not recommended to enter more than two decimal places. | 0,01…1,00 | 0,25 |
| - 0,01 – the event appears immediately when BBox appears in the recognition zone | |||
| - 1,00 – the event appears only when BBox is completely overlapped with the recognition zone |
«Back tracking» handler#
The parameters of the handler are given in Table 44.
Table 44. «Back tracking» handler parameters
| Name | Description | Possible values | Default |
|---|---|---|---|
| Camera | Video stream frame source | Cameras added to the system | - |
| Region for intersection | A drop-down list for selecting the zone with which the vehicle intersections are checked for the scenario to work. | Created zones for the selected camera. The process of creating recognition zones is described in section Recognition zone creation. | - |
| Min threshold intersection | The ratio of the zone of the BBox of the vehicle or the LP that crossed the Region for intersection to the non-crossed one. It is not recommended to enter more than three decimal places. | 0,01…1,00 | - |
| - 0,01 – the event appears immediately when BBox appears in the Region for intersection | |||
| - 1,00 – the event appears only when BBox is completely overlapped with Region for intersection | |||
| Count of bestshots | The minimum number of frames of the current track to select the best frame. When the minimum is reached, the probability of correct recognition of all attributes and symbols of the LP is higher. If there are fewer frames in the track than the minimum, then the best frame will be selected, but the probability of correct recognition of all attributes and symbols of the LP is lower. The more frames are collected to select the best, the more accurate the recognition of vehicle attributes will be. An increase in the number of frames can lead to a decrease in performance. | 1…1000000 | 1 |
| - 1 – high probability of getting incorrect recognition results | |||
| - 1000000 – the time to select the best frame increases, but the quality is higher | |||
| Max count of bestshots | The maximum number of best frames for one track. When the maximum is reached, the selection of the best frames is completed. | 1…1000000 | 2 |
| - 1 – high probability of getting incorrect recognition results | |||
| - 1000000 – the time to select the best frame increases, but the quality is higher | |||
| Bestshot send periodicity | Frequency of sending bestshots | 1000…1000000 | 2000 |
| - 0 – no delay | |||
| - 1…1000000 – pause in ms | |||
| Use track area | Using an additional Detection Zone to work with a part of the road (traffic lane, entrance/exit, stop, etc.). Forming a list of the best frames will only be available for vehicles that overlap the traffic area with a specified threshold. | On/Off | Off |
| Track area | The detection zone with which the vehicle intersections are checked. The introduction of an additional zone will allow you to cut off events that do not need to be recorded. | Created zones for the selected camera. The process of creating zones is described in section Recognition zone creation. | - |
| Track area threshold | The relative area of the Recognition Zone overlapped by the BBox of the vehicle, when exceeding the threshold of which the vehicle will be taken into account to form the best frames. The higher the threshold, the more accurate the detection of the vehicle and LP will be, but the number of passes will also increase. It is not recommended to enter more than two decimal places. | 0,01…1,00 | 0,25 |
| - 0,01 – the event appears immediately when BBox appears in the recognition zone | |||
| - 1,00 – the event appears only when BBox is completely overlapped with the recognition zone |
«Back crawler» handler#
The parameters of the handler are given in Table 45.
Table 45. «Back crawler» handler parameters
| Name | Description | Possible values | Default |
|---|---|---|---|
| Camera | Video stream frame source | Cameras added to the system | - |
| Region for intersection | A drop-down list for selecting the zone with which the vehicle intersections are checked for the scenario to work. | Created zones for the selected camera. The process of creating recognition zones is described in section Recognition zone creation. | - |
| Min threshold intersection | The ratio of the zone of the BBox of the vehicle or the LP that crossed the Region for intersection to the non-crossed one. It is not recommended to enter more than three decimal places. | 0,01…1,00 | - |
«Side crawler» handler#
The parameters of the handler are given in Table 46.
Table 46. «Side crawler» handler parameters
| Name | Description | Possible values | Default |
|---|---|---|---|
| Camera | Video stream frame source | Cameras added to the system | - |
| Region for intersection | A drop-down list for selecting the zone with which the vehicle intersections are checked for the scenario to work. | Created zones for the selected camera. The process of creating recognition zones is described in section Recognition zone creation. | - |
| Min threshold intersection | The ratio of the zone of the BBox of the vehicle or the LP that crossed the Region for intersection to the non-crossed one. It is not recommended to enter more than three decimal places. | 0,01…1,00 | - |
Connecting and configuring cameras#
Cameras are connected and configured in the «Cameras» section, see «Cameras» section»
For correct operating mode «Super Event» it is necessary to connect one mandatory Front tracking camera and N - additional cameras located on the target object.
Let's consider examples of setting up four cameras for operating mode «Super Event».
Setting the Front tracking camera#
Add a camera with the source type RTSP stream. Camera connection parameters (Figure 99).
Load the camera preview from a file or from the CARS Stream.
Add 3 recognition zones. Configure the recognition zones (Figure 100).
A description of the recognition zones is given in Table 47.
Table 47. Description of recognition zones
| Zone | Description | Purpose |
|---|---|---|
| droi | Recognition zone | Main vehicle recognition zone |
| zoneA | First control zone | Determination of the direction of movement of the vehicle |
| zoneB | Second control zone | Determination of the direction of movement of the vehicle |
For a description of the settings and connection options, see «Cameras» section
Setting the Back tracking camera#
Add a camera with the source type RTSP stream. Camera connection parameters (Figure 101).
Load the camera preview from a file or from the CARS Stream.
Add 2 recognition zones. Configure the recognition zones (Figure 102).
A description of the recognition zones is given in Table 48.
Table 48. Description of recognition zones
| Zone | Description | Purpose |
|---|---|---|
| droi | Recognition zone | Main vehicle recognition zone |
| track | Movement area | The area where the vehicle movement is tracked |
For a description of the settings and connection options, see «Cameras» section
Setting the Back crawler camera#
Add a camera with the source type ANPR camera. Camera connection parameters (Figure 103).
Load the camera preview from a file or from the CARS Stream.
Add 1 recognition zone. Configure the recognition zone (Figure 104).
A description of the recognition zones is given in Table 49.
Table 49. Description of recognition zones
| Zone | Description | Purpose |
|---|---|---|
| droi | Recognition zone | Main vehicle recognition zone |
For a description of the settings and connection options, see «Cameras» section
Setting the Side crawler camera#
Add a camera with the source type ANPR camera. Camera connection parameters (Figure 105).
Load the camera preview from a file or from the CARS Stream.
Add 1 recognition zone. Configure the recognition zone (Figure 106).
Description of recognition zones zones is given in Table 50.
Table 50. Description of recognition zones
| Zone | Description | Purpose |
|---|---|---|
| droi | Recognition zone | Main vehicle recognition zone |
For a description of the settings and connection options, see «Cameras» section
«Advanced events» section#
Events of vehicle detection and recognition of vehicle and license plate attributes are displayed in the «Advanced events» section, see «Events» section.
The event filter is currently not supported in the «Advanced events» section
«Advanced events» card#
The event card shows extended information about the event. An example of an event card (Figure 107).
The example event card image shows irrelevant images of the vehicle due to the physical absence of additional cameras in other perspectives.
A description of the information on the event card is given in Table 51.
Table 51. Event card description
| Section | Name | Description |
|---|---|---|
| General | ID | Event ID |
| Camera | The name of the camera that recorded the event | |
| Date | Date and time of event registration | |
| Direction | The direction of movement of the vehicle | |
| Scenario | Scenario of the event registration | |
| Scenario type | The type of the recorded event is set using the scenario. | |
| Incident type | The type of the detected object. | |
| Handler | The name of the scenario handler. | |
| Track ID | Track ID on which the event was recorded | |
| Frame ID | Frame ID with vehicle image | |
| Comment | A link that opens a comment entry window when clicked. After adding a comment, the link changes to the text of the comment | |
| Frame | Track frames | Vehicle images in original (maximum) resolution: |
| - Track start frame – the first frame on which the vehicle was detected; | ||
| - Bestshot – frame with a best shot of the track. | ||
| Front view | Front view of the object | |
| Rear view | Rear view of the object | |
| Side view | Side view of the object | |
| LP on the front | Front license plate image | |
| LP on the rear | Rear license plate image | |
| Additional information | Type | Vehicle attribute showing the recognized vehicle type |
| Brand | Vehicle attribute showing the recognized vehicle brand | |
| Model | Vehicle attribute showing the recognized vehicle model | |
| Emergency services | Attribute of belonging of the vehicle to emergency services | |
| Color | Vehicle attribute showing the recognized vehicle color | |
| LP on the front | License plate attribute, which provides information about the vehicle's front LP and the country of LP ownership | |
| LP on the rear | License plate attribute, which provides information about the vehicle's rear LP and the country of LP ownership | |
| Number of axles | Vehicle attribute showing the recognized number of vehicle axles | |
| Public transport | Attribute of belonging of the vehicle to public transport | |
| Special transport | Attribute of belonging of the vehicle to special transport |
The numbers for each attribute are an estimation of recognition. The score is defined in the range from 0 to 1, the closer the value is to 1, the more accurate the attribute recognition result is.