Will AI-cameras check rear-seat passenger seatbelts, lane discipline? Top FAQs answered

Artificial Intelligence (AI)-enabled cameras have started monitoring traffic rule violations on Kerala roads. The state Motor Vehicle Department (MVD) has installed 726 such cameras in a bid to curb road accidents.

Besides these 726 cameras, the MVD has also deployed four mobile camera units in different parts of the state. Though the cameras have become operational, the MVD will start penalising traffic violators only from May 19.

The MVD's initiative has sparked several doubts. Transport Commissioner S Sreejith responds to questions that have been frequently raised.

The central government revised speed limits through a notification in 2018. But the AI cameras in Kerala are calibrated by the 2014 state notification.

The Centre fixed the maximum speed limit on roads. Each state has the authority to implement speed limits based on respective road conditions. Motorists in Kerala should adhere to the speed limit specified in the state's 2014 notification.

Increasing the speed limits on the proposed six-lane roads is under the state government's consideration.

Where are the cameras installed?

The cameras are installed at major accident-prone stretches and areas where traffic rules are frequently violated. Besides the MVD's AI-enabled cameras, the police will soon be launching its camera network under its Intelligent Traffic Enforcement System.

A meeting chaired by the Home Secretary has decided to interlink the camera networks.

Of the MVD's 726 cameras, 675 are meant to detect seatbelt and helmet rule violations. They will also capture images of two-wheelers carrying more than two riders.

Another eight cameras will monitor overspeeding, while 18 others installed in junctions are Red Light Violation Detection (RLVD) cameras. Illegal parking will be captured by 25 other cameras.

What is the capacity of the system? How much footage could it capture and how long could it be stored?

Any number of photographs could be stored for any length of time. Their backup, too, could be accessed anytime. Once a case is registered and challan generated based on a particular visual, it will be stored until the violator pays the penalty.

If a family is riding a two-wheeler, can they take children with them?

If there are more than 2 people riding a scooter or motorbike, the cameras will capture the visual. If a child is riding with an adult, the minor should mandatorily wear a helmet.

What about rear-seat car riders? Will the cameras capture their image if they have not buckled up?

The cameras won't detect rear-seat riders in the first phase.

In case a motorist touches his/her ear while driving, will the camera mistake it as using a mobile phone?

The cameras will primarily be detecting unseatbelted and helmetless riders. It won't report mobile phone usage if one touches the ear. Currently, the facility to detect people using the hands-free system in cars is not available. It is better not to use the mobile phone while driving since it tends to distract the driver.

Are the cameras capable of finding whether a driver/passenger is wearing the seatbelt if s/he is dressed in black or using a shawl?

Colours do not matter. Different exposures are employed to click the photographs. Scrutiny of photographs will reveal whether a person has buckled up irrespective of the colour or type of dress s/he is wearing.

Will the AI cameras check lane discipline also?

Though AI cameras could be used to detect lane violations, they won't be doing so in the first phase.

Will the visuals captured through AI cameras be shared with the police and other agencies?

Yes. It has been decided that visuals will be shared with other agencies. However, since the cameras are not for surveillance, they cannot follow those fleeing after committing a crime. Additionally, it doesn't have video-recording capabilities. It is meant only to detect traffic violations.

What is the intention behind installing AI cameras? Are they meant to increase the state's revenues by imposing fines on traffic violators?

No, they are not meant to impose penalties. The intention is to keep the people safe. We are willing to reveal the locations where the cameras are installed. At least people will adhere to the rules at such locations. The imposition of fines is to create awareness. The department is dreaming of a time when no one would have to pay penalties.

What is the estimated number of violations are you expecting to capture daily through the AI cameras?

We installed 81 cameras at various locations in Thiruvananthapuram during the trial run. The cameras reported around 50,000 violations a day. If we take the numbers from cameras placed across the state, the number would exponentially go up.

The cameras are of good quality with night-vision capabilities. Clear images could be captured using infrared flash at night.

Initially, the cameras will be capturing images of those riding without seatbelts and helmets, overloading on two-wheelers (more than two riders), illegal parking, overspeeding, using a mobile phone while driving, and jumping signals at junctions.

Software in the control room will automatically generate a challan after finding the vehicle's registration number. A message will alert the registered owner of the vehicle once the challan is generated. The district-level control rooms will initiate further action on the information passed down by the control room at the state level.

Signboards and markings are missing on several roads?

Different agencies, such as the National Highways Authority of India, the Public Works Department, and Local-self government bodies are responsible for the upkeep of roads. The Road Safety Authority (RAS) has instructed these agencies to ensure signboards.

The RSA and Keltron (the implementing agency) will put up signboards at locations where the cameras have been installed.

Will VIP vehicles be exempted?

Emergency vehicles such as ambulances, fire tenders, police vehicles, etc., that have beacon lights have been exempted. There are no other relaxations in the rules.