Safety of women and children has become an agenda for the government, especially in the Nirbhaya case. In 2019, the Ministry of Home Affairs announced the Safe City Project under the Nirbhaya Fund Scheme and allocated a budget of approximately INR 2919.55 crore.

The project aims to make women safe, secure and empowered in eight major Indian cities — Delhi, Kolkata, Mumbai, Chennai, Bangalore, Hyderabad, Ahmedabad and Lucknow. While the central government has fully funded the project in Delhi, in other cities the center will fund 60 percent of the project.

“These projects include identifying hotspots for crimes against women and building capacity in the community through various components, including infrastructure, introduction of technology and awareness programmes,” the ministry said.

Bangalore Safe City Project

In Bangalore, the Safe City project will be implemented in phases and Honeywell Automation has won the tender to commission and maintain the security infrastructure.

The Bangalore City Police is already equipped with around 1,100 CCTV cameras for surveillance. As part of the project, about 7,500 cameras will be installed at 3,000 locations in the city. About 6,300 of these will be stationary cameras, 800 pan-tilt-zoom cameras and 400 high-definition cameras.

(Source: Twitter)

The tender lists 2,744 locations where new cameras should be installed. The number of cameras offered for each location varies from one to nine, and areas such as important traffic intersections, market areas, deserted streets, educational institutions, offices and transport hubs are highlighted.

Among the eight cities selected by the government for the Safe City project, Bangalore is the first to submit this Request for Proposal (RFP). “I think Bangalore Police has seen how AI-powered technologies can help keep women and children safe, how AI can help the common citizen and make the city of Bangalore safer and more secure. That was the whole point of the project,” said Yatin Kavishwar, Founder and CEO of Awiros.

Using AI-based video analytics

The use of video analytics by law enforcement agencies in India is nothing new. In 2019, apart from CCTV cameras, the UP Police has been cracking down on autoloaders through video analytics and database analysis. However, this may be the first time that video analytics has been used on such a large scale by law enforcement agencies.

“I think the Bangalore Safe City project has paved the way for widespread adoption of video AI in government, because you need a path-breaking RFP like that, someone to take a pole position and have the guts to embrace a new technology. horizon”.

“AI has enabled the police and the administration in general to have a unified view of the scenario across the state. Earlier, what happened was that all these solutions were placed in silos; it was at the camera level, not even at the network level,” said Yatin Kavishwar.

Bangalore Police uses AI-powered video analytics service provided by Awiros for number recognition, facial recognition, area tracking for crimes against women, road health, detection of encroachments and crowding and more uses

This enables the Bangalore City Police to monitor and analyze the situation in real time and take timely action against any crime that may be committed by the offenders for the safety of the citizens.

The video feed from all cameras is sent to the data center. A Command Control Center (CCC) will be set up with live video feed. The command center uses video analytics solutions to identify a vehicle or person involved in a crime during an emergency.

“The platform will definitely increase the speed of decision-making for the police. I think this is one of the important aspects of the project,” said Yatin Kavishvar.

Automatic number detection

The license plate recognition application by Awiros is a highly accurate detection system that can search thousands of live feeds and track a vehicle’s journey from just a few attributes (such as the last two digits of a license plate) and without any humans. intervention.

“A vehicle is often involved in a crime. So, using multiple attributes of the vehicle, the Bangalore police can identify what the footprint of that vehicle is,” said Yatin Kavishwar.

“The learning pipeline is managed by the Annotation, Training, and Orchestration Module (ATOM). So this is our homegrown module and this data never leaves the platform. What ATOM does is basically two things. First, it is a data collection tool. Now, when these algorithms run on the platform, a lot of metadata is generated, and most of it is in the form of photos.

In the case of a newly launched SUV model, Yatin Kavishwar says that the algorithm may not be trained on the latest models available in the market. “Now a batch of such photos is being created on the ATOM platform. This batch is then sent to the annotation group. There is an annotation team sitting on the customer’s premises. The annotated images are then used to train algorithms in ATOM, which are then fed back into the Awiros Video AI platform.

Facial recognition system

AI-powered video analytics will also help the Bangalore police in tracking individuals. According to the tender document, facial recognition systems will be used to identify and track individuals involved in stalking women or other illegal activities in public spaces.

The CCC relays alerts from the video analytics system to on-site police. At the same time, police officers working in the field will send the photos of the suspect to the NSC for checking against the criminal database.

However, the use of facial recognition systems by law enforcement agencies has raised concerns, especially since the country does not have a data protection law. Another concern with using FRT is that it can be used for surveillance.

“We are tasked with identifying criminals from the database. Now this database comes from Bangalore police or any police that we work with. It is part of the State Criminal Records Bureau database or part of the National Criminal Records Database. So for us, we are not profiling individuals; Instead, we are only targeting the face of the person in the database of criminals,” said Yatin Kavishwar.

Source link