Latest news with #MaharashtraResearchandVigilanceforEnhancedLawEnforcement
News18
3 days ago
- News18
Maharashtra's 'MARVEL': Police Solve Hit-And-Run Case Within 36 Hours Using AI
Last Updated: Authorities highlighted that while reviewing 12 hours of footage manually could take days, AI-powered MARVEL completed the analysis in as little as 12–15 minutes. In a remarkable example of tech-aided policing, the Maharashtra Police employed artificial intelligence to solve a horrific hit-and-run case in Nagpur within just 36 hours. The incident occurred on August 9 (Raksha Bandhan), when a speeding truck struck a motorcycle, fatally injuring the pillion rider. In a distressing turn, the husband tied his wife's body to his bike and attempted to take her home, a moment captured in a video that quickly went viral. The only clue available to investigators was that the truck bore red markings and nothing else — no make, model, or distinguishing features. Harssh Poddar, Superintendent of Police (Nagpur Rural), told NDTV that the victim provided 'very minimal information." Faced with such scant detail, the police turned to MARVEL (Maharashtra Research and Vigilance for Enhanced Law Enforcement), the state's first AI-powered law enforcement tool. According to NDTV, the team analysed hours of CCTV footage from three toll nakas within 15–20 km of the crash site using two separate computer vision algorithms. The first algorithm filtered trucks with red markings; the second detected anomalies in average speed — signaling potential involvement. Within minutes, they pinpointed the suspicious vehicle, which led them to track and arrest the driver 700 km away on the Gwalior–Kanpur highway. The tool — developed collaboratively by the Maharashtra government, IIM Nagpur, and Pinaca Technologies — underscores the potential of data-driven policing. It is intended to aid in missing-person cases, fatal accidents, vehicle thefts, organized crime, and cyber fraud — not just reactive investigations but also predictive interventions. Recent Similar AI-Aided Police Successes In Kerala, AI helped crack a 19-year-old triple murder case. Police used digitally aged photographs of the suspects, matched them against social media, and made arrests in early 2025. In February 2006, a horrific crime in Anchal, Kerala, saw Ranjini and her 17-day-old twin daughters brutally murdered. The prime suspects, Army personnel Divil Kumar and Rajesh, fled and vanished for 19 years. In January 2025, Kerala Police used AI to digitally age their old photos, matching them with social media images, including a wedding picture. This breakthrough led to their arrest in Puducherry, where they lived under false identities. The CBI confirmed their involvement, finally closing a nearly two-decade-old case. view comments First Published: Disclaimer: Comments reflect users' views, not News18's. Please keep discussions respectful and constructive. Abusive, defamatory, or illegal comments will be removed. News18 may disable any comment at its discretion. By posting, you agree to our Terms of Use and Privacy Policy.
Indian Express
3 days ago
- Indian Express
Man carries wife's body on bike: How Nagpur Police used AI to find truck involved in fatal accident
In a breakthrough, the Deolapar police in Maharashtra's Nagpur district have successfully used artificial intelligence-based analysis to identify the vehicle involved in a fatal accident after which the victim's body was carried on a motorcycle by her husband. Gyarasi Amit Yadav, 30, died when an unidentified vehicle hit the motorcycle she was riding pillion on with her husband Amit Bhura Yadav, a resident of Lonara at Gumti in Nagpur district, in the Morfata Dargah area on August 10. A video of Amit carrying his wife's body on his motorcycle after the accident was widely shared on social media platforms. Based on Amit's complaint, the police registered a case on August 11 on charges of rash driving and causing death by negligence under Bharatiya Nyaya Sanhita sections 281 () and 106(1) along with Motor Vehicles Act sections 184 and 134(b). With no initial leads on the accused or the vehicle, investigators turned to the AI-powered system MARVEL (Maharashtra Research and Vigilance for Enhanced Law Enforcement) under the guidance of Harssh A Poddar, Superintendent of Police, Nagpur Rural, the police said. The system analysed CCTV footage and traced the vehicle involved—a red Tata truck (UP 14 MT 2190). HOW AI WORKED Harssh A Poddar told The Indian Express that the police used AI algorithms developed at MARVEL to analyse four hours of footage from three toll nakas. While the first algorithm sieved out all the red trucks, the second algorithm did a speed-based calculation of these trucks to identify which truck could be involved. On the basis of the AI alert generated, the police arrested the accused yesterday from 700 km away near Kanpur. Explaining that they relied on two independent algorithms to crack this case, Poddar said that the first one was based on computer vision. 'The only clue we had from the victim was that the truck had red markings. Normally, going through four hours of CCTV footage from three toll nakas would take ages of manual effort. Now, the system scans the footage in just 15 minutes and picks out all vehicles that match the description, in this case, trucks with red markings or a red coloured truck.' He added, 'The second algorithm looks at CCTV feeds from different toll nakas and calculates the average speed of vehicles. If there's a mismatch between the expected travel time and the actual time recorded, it signals an anomaly. That anomaly could mean the vehicle met with an accident or stopped somewhere. In this case, there were no petrol pumps on that stretch, so the stop pointed us towards an incident.' 'By combining these two independent datasets, red markings and unusual speed patterns, we narrowed it down to a specific truck on the Chandrapur-Delhi highway. We finally intercepted it between Delhi and Kanpur,' Poddar said. The police added that the driver, from Farrukhabad, had no prior criminal record. It was the precise triangulation of these two systems that led the police straight to him. On August 16, police teams tracked down and arrested the driver of the truck, Satyapal Rajendra, 28, a resident of Nagla Kharik village in Mahoi in Farrukhabad, Uttar Pradesh. They also seized the truck. MARVEL is a special purpose vehicle (SPV) owned by the Maharashtra government, it has been created in collaboration with the Indian Institute of Management (IIM) Nagpur. The Maharashtra government, IIM Nagpur and Chennai-based Pinaca Technologies entered a tripartite agreement on March 22 in 2024 and the company was registered under The Companies Act, 2013. MARVEL has been created to enhance intelligence capabilities and improve crime prediction and solving, enabling the state police force to tackle law enforcement more effectively using AI. Maharashtra has become the first state in the country to create an independent entity for AI in law enforcement. MARVEL also assists in addressing various crimes, including cases related to missing women and children, fatal accidents, stolen vehicles, and organised gang-related activities, to name a few. It will also assist the cyber cell tackle fraud cases.
NDTV
3 days ago
- NDTV
How AI Helped Maharashtra Cops Crack Hit-And-Run Case Within 36 Hours
Nagpur: The man was notably short on specifics. All he knew was that the truck that ran over his wife in Nagpur had red markings on it. He could neither identify its size nor make. The police used artificial intelligence or AI, to help arrest the accused within 36 hours. On Raksha Bandhan, August 9, the man and his wife were on a bike when they were hit by a speeding truck in Nagpur. The woman fell on the road. The truck ran her over and fled the scene. The man then tied his wife's body to his bike - a disturbing video of the incident is in wide circulation - to take her to their village in Madhya Pradesh. The victim was able to provide "very minimal information", Harsh Poddar, Superintendent of Police, Nagpur Rural, told NDTV. Asked how the cops zeroed in on the accused with so little information available, Mr Poddar said they collected the CCTV data and ran it through artificial intelligence algorithms. "What was done here was that CCTV footage, or as we call it metadata, of three separate toll nakas, which were at a distance of 15-20 km from each other, was collected and it was analysed using two AI algorithms. Both of which were based on a technology called Computer Visual," Mr Poddar said. "The first algorithm, what it did was it analysed this total CCTV footage to pull out and identify trucks with red markings. The second one analysed the average speed of all these trucks to be able to give us an alert as to which truck could be involved. On that basis, one truck was identified. Day before yesterday, a team from Nagpur rural police arrested the accused and seized the truck from the Gwalior-Kanpur highway, which is about 700 km from Nagpur. And we have now arrested the accused and the case has been cracked within 36 hours using AI," the officer elaborated. Maharashtra has created a special purpose vehicle named MARVEL (Maharashtra Research and Vigilance for Enhanced Law Enforcement), the first state-level police AI system in the country. It is owned wholly by the state. "It has been created to introduce and use AI in law enforcement agencies as well as other departments of the government of Maharashtra. The basic idea is to be able to analyse and use government metadata to be able to create AI solutions that belong to the government of Maharashtra instead of getting them from outside," said the police officer. "We have been working with as many as 13 different departments to create solutions. The main fuel for AI is data and given the government is such a large repository of data we have been analysing it to create AI solutions across the entire ecosystem. I am also happy to say that many other large states have expressed interest in this model since Maharashtra is the first state that has created this and are in the process of replicating Marvel," he said further. Asked how physical investigation is different from AI, the officer said the use of technology helps speed things up. The data could be analysed within minutes to help arrive at solutions, said the police officer. "We get several terabytes of metadata, these could be complex financial crimes, organised crimes where large amounts of telecom data, for eg, is there. The earlier model was to use experienced police officials who would analyse patterns in this data. There was obviously a big margin for human error in such cases. And it could take several weeks, to months to analyse large sets of data. Today, with the power of AI and high-speed processors, we are able to analyse this data within a matter of minutes. So for example, 12 hours of CCTV data that existed in this case, could be analysed within 12-15 minutes, which earlier could have taken more than a day," said the officer.
Indian Express
13-07-2025
- Indian Express
Maharashtra deploys AI-powered MARVEL system for wildlife tracking and surveillance system
In a significant step towards reducing human-wildlife conflict in Maharashtra, the government-led initiative MARVEL (Maharashtra Research and Vigilance for Enhanced Law Enforcement) has signed an MoU with the state forest department. The technology to be developed under this project is being adapted to forest-edge zones near tiger reserves like Tadoba, Nawegaon-Nagzira and Pench, as well as proposed areas like Ramtek, Deolapar, Parashivni, and Pauni. MARVEL was initially developed for law enforcement surveillance and is harnessing its AI-driven capabilities in governance-related interventions. In the case of the forest department, the project aims to detect and deter wildlife from entering villages, ensuring both human safety and animal conservation using AI capabilities. The system deploys smart camera poles equipped with AI modules near forest boundaries. These function as 'virtual walls,' automatically detecting approaching animals and triggering gentle deterrents, such as lights and sounds, to nudge them back into the forest. This soft, non-intrusive method is proving to be a practical solution to mitigate conflict without harming the animals. To address the large number of innocent lives being lost to man-animal conflict, a special Memorandum of Understanding (MoU) was signed on Saturday, in Nagpur, with MARVEL for remedial measures. The agreement was exchanged in the presence of Revenue Minister and Guardian Minister Chandrashekhar Bawankule, Minister of State for Finance Ashish Jaiswal, CEO MARVEL, and district superintendent of police (rural) Dr Harssha Poddar, and senior officials of the Tiger Projects. MARVEL was established with the foresight of Chief Minister Devendra Fadnavis to effectively use artificial intelligence (AI) to reduce accidents and crime. Under this MoU, a total of 3,150 AI-based cameras will be installed along the borders of villages surrounding the tiger projects. Sirens will be set up in densely populated areas of the villages and connected wirelessly to these cameras. The cameras will have technology to independently identify tigers and leopards using a combination of visual, thermal, and acoustic data. As soon as a tiger or leopard is captured by a camera, the connected sirens in the village will alert the villagers. As per the agreement, 875 cameras and sirens will be connected in the Pench Tiger Project, 525 in the Tadoba-Andhari Tiger Reserve, where 140 pre-installed cameras will be upgraded with audio modules, 600 in the Navegaon-Nagzira Tiger Reserve, and 1,145 units in the Nagpur Forest Division (Territorial), said Dr Harssha Poddar. The camera systems are equipped with high-resolution sensors, night vision, and built-in microphones. They primarily use edge-based AI model inference for real-time detection, reducing latency and dependency on network connectivity, with cloud support for analytics, archival, and remote management. For better detection level alerts, we prefer to use a cloud or server, said MARVEL officials. Real-time alerts are sent through multiple channels, including SMS, mobile apps, Telegram, and local hooters. These alerts are available in regional languages and typically reach stakeholders within seconds, due to the system's sub-second processing capability. Pilot deployments near Tadoba have already shown positive outcomes, informed the officials. Human-wildlife interactions have decreased, and no fatal incidents have been reported in the areas covered since MARVEL's implementation. Comprehensive training modules, field workshops, and remote support have been planned to ensure local staff are skilled in operating, maintaining, and expanding the AI systems. The expansion roadmap follows a modular strategy, with each deployment phase lasting approximately 24 weeks. Initial efforts in this project focus on high-risk buffer villages, with further scale-up planned in a phase-wise manner over the next 12-18 months, depending on the outcomes. Looking ahead, MARVEL aims to enhance its capabilities further with drone surveillance, AI-driven path prediction, and deeper analytics. Thermal imaging is already a core component, and future integrations will be pursued through partnerships with technology providers and stakeholders. The initiative is backed by a strong collaborative framework in the form of a Special Purpose Vehicle (SPV). MARVEL operates under the Government of Maharashtra, in partnership with IIM Nagpur for academic and policy collaboration, and with Pinaca Technologies as the technical implementation partner.
