13
IoT-Enabled Vision System for Detection of Tint Level
than 35% VTL color percentages. For the second set of input images, the RGB percentages are 14%, 16%, and 16%, respectively, which are less than 35% of VTL, and hence
one can conclude that the dark tinting is present in the windows of those vehicles. The
various input images given clearly identify and classify the vehicles based on tinting
levels.
In conclusion, in the present work, we extract the relevant, meaningful images from the
captured video frames. From the captured images, through motion detection techniques,
the presence or absence of a vehicle is determined. If a vehicle is present, it is classified using appropriate imaging techniques. Depending on the type of vehicle, the ROI is
identified and vehicle image alone is cropped for further processing. Depending on the
vehicle type, the windshield/window region is identified and that portion of the image
is used for further processing. Using three different techniques, namely color segmentation, contour detection, and histogram analysis, the tint level of windshield/window
region is estimated. The tint level on the detected vehicle is verified against the database
of government permissible limits. If the tint level exceeds government norms, the imaging system extracts the license plate of the vehicle. If the level exceeds the limits of regulation bodies, the controller/processor interfaced with the surveillance system extracts
the GPS details from the GPS receiver and appends it with date and time. A messaging
system (such as SMS/MMS/e-mail) interfaced with the surveillance system automatically
generates an evidence consisting of latitude, longitude, date, time, license/plate image,
vehicle with tinted window/windshield image for issuing necessary tickets. The messaging system stores the evidence at the central server. At the server side, using the number
Vehicle under
surveillance
Tint detection
module
Number plate
detection module
Issue token
Client machine
Client
Network
Server machine
Data storage
Server
GPS
Date and
time
Controller
Location
ID
FIGURE 1.10
The controller modules for issuing ticket with evidence.
IoT-Enabled Vision System for Detection of Tint Level
than 35% VTL color percentages. For the second set of input images, the RGB percentages are 14%, 16%, and 16%, respectively, which are less than 35% of VTL, and hence
one can conclude that the dark tinting is present in the windows of those vehicles. The
various input images given clearly identify and classify the vehicles based on tinting
levels.
In conclusion, in the present work, we extract the relevant, meaningful images from the
captured video frames. From the captured images, through motion detection techniques,
the presence or absence of a vehicle is determined. If a vehicle is present, it is classified using appropriate imaging techniques. Depending on the type of vehicle, the ROI is
identified and vehicle image alone is cropped for further processing. Depending on the
vehicle type, the windshield/window region is identified and that portion of the image
is used for further processing. Using three different techniques, namely color segmentation, contour detection, and histogram analysis, the tint level of windshield/window
region is estimated. The tint level on the detected vehicle is verified against the database
of government permissible limits. If the tint level exceeds government norms, the imaging system extracts the license plate of the vehicle. If the level exceeds the limits of regulation bodies, the controller/processor interfaced with the surveillance system extracts
the GPS details from the GPS receiver and appends it with date and time. A messaging
system (such as SMS/MMS/e-mail) interfaced with the surveillance system automatically
generates an evidence consisting of latitude, longitude, date, time, license/plate image,
vehicle with tinted window/windshield image for issuing necessary tickets. The messaging system stores the evidence at the central server. At the server side, using the number
Vehicle under
surveillance
Tint detection
module
Number plate
detection module
Issue token
Client machine
Client
Network
Server machine
Data storage
Server
GPS
Date and
time
Controller
Location
ID
FIGURE 1.10
The controller modules for issuing ticket with evidence.