The rapid development of artificial intelligence in the electronic security market has facilitated access to a greater volume of information on the scene thanks to the classification of objects of interest (mainly vehicles and people).
The technological evolution of license plate recognition cameras has also allowed greater precision in the capture of license plates at speed, which until recently, was unimaginable at a reasonable cost.
This has provided access to a new era in smart cities, thanks to these devices that allowed operators with much more information to gain in depth knowledge of the real-time behaviour of their municipality, as well as the possibility of solving complex situations through recorded video.
The SGSE Forensic LPR at Milestone
SGSE has developed an integration in Milestone to carry out forensic analysis on the recorded video with the information from the artificial intelligence of each certified license plate recognition camera in this integration.
With the Forensic LPR, in any of the software editions of Milestone, we can:
Search the recorded video by full or partial license plate.
Search the recorded video by LPR camera characteristics integrated into the Forensic LPR.
Create blacklists and whitelists.
Create PDF reports.
Create CSV reports.
The available characteristics are based on each camera, the most common being:
License plate.
Type of vehicle.
Colour of vehicle.
Type of registration.
Vehicle make.
Country of registration.
When the operator needs to open an investigation into an event, they can find all the search tools in the Search tab of Milestone Smart Client shown in the image below to work more efficiently in a single place.
Within the Search tab, the operator can search by:
Movement.
Alarms.
Events.
Markers.
Metadata.
To start a search, we must select a time window within the different predefined options that appear in the drop-down menu shown in the following image:
Or customize the search through the calendar:
Subsequently, we select the camera or group of cameras on which we want to perform the search:
Once the search on the camera or the set of cameras has been executed, we have the total number of results (in our example below, a total of 842) distributed in the time window that we have defined over which the operator can move easily.
By clicking on any of the video thumbnails obtained from the search, a blue square will appear, indicating that we have opened an investigation on this thumbnail through its recorded video sequence in the right margin of Smart Client.
In our case, we see it in the video segment recorded from camera KM47 for the moments before and after the search timeline to access all the details.
This way we access all the details, as well as having the location of the camera.
We can finish the investigation by exporting the recorded video sequence as shown in the image below:
Filters for intelligent searches on video recorded with the Forensic LPR
Thanks to the Forensic LPR developed by SGSE we can find a significant number of additional filters for intelligent searches.
This allows us to provide the operator with a powerful forensic tool to solve incidents in our municipalities.
Within the Search tab, previously explained, we will have different options to run the search in the recorded video, among which is the LPR Search module which gives access to the filters of the different characteristics provided by this LPR camera model.
Once we click on Search LPR, we see the different search filters such as the License Plate Type shown below:
We can filter according to the type of license plate and have the option of searching for those vehicles that do not have license plates within the LPR camera’s field of view as well as vehicles that transport dangerous goods.
The following set of images details these different types of vehicles, which can be filtered based on their license plate type.
We can continue in the filter options with the Vehicle Make within the LPR camera’s field of view, as detailed below.
A total of 86 vehicle makes are available, shown in the table below.
Hyundai
SsangYong
Alfa Romeo
Toyota
Citroen
Great Wall
KIA
Fiat
Infiniti
Volkswagen
Scania
Smart
Benz
MAN
Maxus
Nissan
Volvo
JAC
Ford
Lexus
Jaguar
Isuzu
Seat
GMC
BMW
Land Rover
Lincoln
Chevrolet
Daihatsu
JMC
Mitsubishi
Dongwo
SAAB
Renault
Subaru
FAW
Opel
Iveco
Yutong
Suzuki
MINI
Guangzhou Yunbao
Skoda
JEEP
Joylong
Daewoo
Porsche
Geely
Audi
Chery
Cadillac
Mazda
Dodge
JBC
GAC HINO
Chrysler
An´Kai
Peugeot
Acura
Haima
Foton
King Long
Dongfeng
Geely-Emgrand
Perodua
UD
BYD
Renault Samsung
Proton
HICOM
Hyundai-Rohens
SsangYong-Old Version
Equus-Old Version
CNHTC
Rolls-Royce
Beiben Truck
Haval
Hino
Kia-Borrego
Chang´an
Alfa
Apart from the make of the vehicle, we can also carry out a search by its Colour, a total of 11 colours being available in an additional filter.
RED
YELLOW
GREEN
BLUE
ROSE
PURPLE
CYAN
BRONW
GREY
WHITE
BLACK
This set of filters can also work in combination with the Vehicle Type for forensic analysis of the recorded video.
Car
Pickup
Van
Bus
Truck
Small truck
Scooter
Motorcycle
SUV
Pedestrian
Thus, the development of SGSE allows us to carry out a precise forensic analysis of the recorded video of our facility in order to resolve incidents quickly and effectively.
Searches for partial license plates in the Forensic LPR
To facilitate the operator’s tasks and be able to solve complicated situations in which we do not have all the information, the Forensic LPR allows searches with partial information of the desired license plates based on a set of incomplete alphanumeric information.
For example, it will be enough to enter a set of digits in the searches so that the Forensic LPR shows us all the different license plates captured with their associated video that contains said digits.
Whitelists and/or blacklists in the Forensic LPR
Along with the previous display related to the forensic tools in the recorded video, there is also the option of raising events in the Forensic LPR, thanks to the creation of white and/or blacklists.
Once the white and/or blacklists have been created as shown in the image below, we can incorporate the license plates of those vehicles that we want to trigger an event when they are captured by certified LPR cameras in the Forensic LPR.
The list of license plates can be easily included by importing a file with the set of information.
These events allow us to trigger actions within Milestone through the rule engine.
So, for example, we can activate the raising of an exit barrier when a license plate belonging to a list of authorised vehicles (whitelist) is read.
Likewise, we can trigger, for example, the sending of an email or the activation of a red light at an exit when a license plate from a list of unauthorised vehicles is read (blacklist).
AUTHORISED REGISTRATION EVENT
UNAUTHORIZED REGISTRATION EVENT
Likewise, we can escalate the events to the alarm category to be managed by the operator.
PDF reports from the Forensic LPR
Thanks to the Forensic LPR, it is possible to obtain detailed PDF reports with all the information linked to a specific vehicle through its license plate, as illustrated by the following image:
CSV reports from the Forensic LPR
All vehicle information (license plate, characteristics, LPR camera that has captured the information and the date and time) can be obtained in CSV format through the Forensic LPR to be processed in subsequent analysis.
LPR cameras certified in the Forensic LPR
The models listed in the table below are currently certified in the Forensic LPR developed by SGSE.
iDS-TCM403BI-0832
iDS-TCV500-BI/1550
iDS-TCV900-BI/1440
iDS-2CD7A46G0/P-IZHSY
HC121@TCR-08S-Z
Environment
Urban and interurban
Brand
Hikvision
Uniview
Sensor
1/1.8″ CMOS
2/3″ GMOS
1″ GMOS
1/1.8″ CMOS
1/2.8″ CMOS
Focal lenght
Motorized 8-32mm
Motorized 15-50mm
Motorized 11-40mm
Motorized 2.8-12mm
Motorized 4.7-47mm
Resolution
4Mpx
5Mpx
9Mpx
4Mpx
2Mpx
Vehicle attributes that can be detected
– License plate – Type of vehicle – Colour (During the day) – Brand – Country of registration
– License plate – Type of vehicle – Colour (During the day) – Brand
– License plate – Type of vehicle – Colour (During the day) – Brand
– Type of vehicle – Country of registration
– Type of vehicle – Colour (During the day)
Compression
H.265 / H.264 / MPEG
IR range
100 m
30 m
50 m
100 m
50 m
API
ONVIF (S,G) e ISAPI
ONVIF (S,G) e ISAPI
ONVIF (S,G) e ISAPI
PROFILE (S, G, T), ISAPI, SDK, ISUP
ONVIF (S,G) e ISAPI
Operating system
Linux
Protection
IP66 / IK10
IP66
IP66
IP67 / IK10
IP67 / IK10
Storage
microSD/TF card 128Gb
microSD/TF card 128Gb
microSD/TF card 128Gb
microSD/TF card 256Gb
microSD/TF card 256Gb
Power supply
12VDC, 24VDC, PoE
24VDC, 100 a 240 VAC
24VDC, 100 a 240 VAC
12VDC, PoE
12VDC, 24VDC, PoE
Embedded LPR
Yes Detection rate = 98% Reading rate = 95%
Yes Detection rate = 98% Reading rate = 98%
Yes Detection rate = 98% Reading rate = 98%
Yes Detection rate = 98% Reading ratea = 98%
Yes Detection rate = 98% Reading rate = 95%
Max. pickup speed
160 Km/h
250 Km/h
250 Km/h
100 Km/h
90 Km/h
Design your electronic security solutions for Smart Cities with SGSE