As automated video-based surveillance and enforcement of traffic offences is catching up in many Indian cities, experts signal the need to streamline the electronic enforcement ecosystem as this will be critical to scale up automation extensively across cities, say experts.
According to the Ministry of Road Transport and Highways statistics, the total number of road accidents have risen from 4,64,900 in the year 2017 to 4,67,000 in 2018 and the total number of deaths due to the crashes increased from 1,47,900 to 1,51,400 during the period.
Traffic rule violations largely go unchecked in India because of manpower and resources constraints while the vehicles hitting the roads continue to increase exponentially. A weak penalty system under the Motor Vehicles Act 1988 has been superseded by the Amendment Act 2019 with stronger penalties and enforcement protocols that addresses deficiency in human dependent enforcement with minimal use of technology.
The electronic enforcement landscape in India
The Motor Vehicles (Amendment) Act 2019 urges states to augment enforcement and regulation of traffic using technology and thereby eliminate human error and corruption. Section 136A of the current Act addresses the disparity that exists in enforcement in different states. Automation initiatives were expected to increase consistency and evidence-backed enforcement, which is lacking among states.
Recent studies have shown that human surveillance proves ineffective; as the duration of monitoring of videos increases, the errors made by humans also increases (IIIT Hyderabad study, â€śDetection of Motorcyclists without Helmet in Videos using Convolutional Neural Networkâ€ť).
â€śThere is a lot that is happening in the technology landscape with regards to traffic. Technologies such as computer vision, IoT and machine learning are being integrated for streamlining traffic management. Notably, China is definitely a leader in this as it has installed numerous CCTVs around the country, which other countries havenâ€™t. This will further allow China to quickly implement technologies to effectively manage traffic. While these technologies are functional, concerns of data privacy is slackening its adoption. However, now governments around the world are finding ways to leverage technologies,â€ť says technology writer Rohit Yadav.
Studies in the effectiveness of speed detection cameras for red light violations have been conducted intensively in the US and support decisions to scale up automated enforcement. According to one such study published by the Centres for Disease Control and Prevention, speed cameras can reduce crashes substantially.
The Integrated Traffic Management System (ITMS) provides greater and accurate information in real time to traffic managers to predict and analyse traffic in order to avoid congestion and gridlocks on the roads, thus making commutersâ€™ life more stress-free.
A wide range of surveillance and enforcement suites of the ITMS have been developed for traffic managers and RTO enforcement personnel. They include: Automatic Number Plate Recognition (ANPR) System, Red Light Violation Detection (RLVD) System, Speed Violation Detection System PA System, E-Challan System, Command Centre and control room along with TMC (traffic management console).
Typically these are powered by AI (artificial intelligence), computer vision, pattern recognition and intelligent video computing platforms. These suites enable storage of the number plate of every vehicle in a database, which can be searched at any later point in time. On detection of any violation, an alert signal is generated at the control room. The software is integrated with the Motor Vehicle Departmentâ€™s website. Subsequently, challans are prepared based on the collected data.
As on date, 164 cities have some level of electronic surveillance to monitor and enforce traffic violations. Considerable differences, however, exist in the level of automation, sophistication and technology, with some components of the system functioning in AI, Deep Learning Architecture, while some work on video analytics based on computer vision and image processing. Some IP cameras require the support of a central network video recorder (NVR) to handle the recording, video and alarm management. Others are able to operate in a decentralised manner without an NVR, as the camera is able to record directly to any local or remote storage media. For data security, private vendors guarantee high security and flexibility to integrate. Nevertheless, many cities are yet to integrate with the National Informatics Centre (NIC) system.
Very few cities have full-fledged ITMS and among them only few have integrated regulation and enforcement components. In Bengaluru, where an advanced Traffic Management Centre with a gird of 11,000 surveillance cameras, adaptive signals and AI-based RLVD and speed detection was set up by the City Traffic Police for the first time in the country four years back, the enforcement components like RLVD and speed detection are not integrated and stand alone. E-challan generation with automatic SMS alerts to the vehicle owner is not yet in place.
In Coimbatore and Chennai, where AI and computer vision powered systems have been installed on a pilot basis, seat belt and helmet detection are done with AI suite in Coimbatore, while in Chennai the two are done with computer vision while speed and RLVD work on the AI suite. There is considerable disparity in the amount of manual interface in the validation and e-challan generation stages among systems installed in various cities that make the system functionality manpower-intensive in many cities. Bhopal Smart City boasts of a fully integrated ITMS with an automatic e-challan generation system.
Why uniformity and standardisation helps
In Delhi and Chhattisgarh, the RLVD system and speed camera installations are integrated with NICâ€™s platform. This means that when a violation is detected by surveillance cameras, e-challans are automatically processed after validation and simultaneously updated in the vehicle ownersâ€™ database at the RTO. But in many other cities like Bengaluru, Chennai, Coimbatore and Vadodara, the system functions independently of NIC; violations are detected automatically but the validation and raising of challans is a manual process.
In Coimbatore and Chennai, where about 12,000 to 15,000 violations are detected every day in the video surveillance zones, with 130 and 62 cameras in 13 and 5 key intersections respectively, hardly 300 to 400 penalty challans are raised due to shortage of manpower. Integration with NIC would enable a high volume of challans per day. A request has been made to onboard the Coimbatore and Chennai systems on the NIC platform, say VK Surendernath, Assistant Commissioner of Police (Traffic), Chennai West Zone, and Raj Kannan, Assistant Commissioner (Traffic), Coimbatore West.
Where the integration is complete with NIC, cities are able to quantify the results, clocking high penalty collections. For example, in Delhi where 100 RLVD and 25 speed cameras were installed since June 2019, 17 lakh on-spot challans have been issued till October 2019 through these devices. Besides this, 37 lakh traffic violation notices were reportedly sent through the e-challan system based on automated video surveillance technology like speed guns and other devices. In Bhopal, where the e-challan suite is integrated with ITMS, a similar high quantification is recorded. In 30 ITMS junctions with more than 300 cameras, about 80 lakh challans have been generated as of October 2019 after the implementation. Revenue collection crossed Rs 1 crore, according to reports.
Joydeep Shome, Senior Technical Director, NIC says that the organisation has developed a common platform ready for any implementing agency to come on board. But they would require quite a large resource for the back-end operations, such as servers and compatible software, he adds. According to him, systems in Delhi, Chandigarh, Agra and Mumbai were integrated with NIC while systems in other cities are functioning independently.
In Coimbatore, it was an NGO floated by local industrialists that had funded the installation. The system in Chennai is supported by the CSR arm of Hyundai Motors. In Delhi, the 25 speed camera system is supported under Maruti Udyogâ€™s CSR initiative. Any upgradation or expansion of the system would require further support from donors as far as Coimbatore is concerned, according to Assistant Commissioner Raj Kannan.
The inconsistencies make any assessment of measurable impact difficult, experts say. Moreover, standardisation is crucial to ensure efficiency, data integrity, security, reliability and transparency. All over India installations of automated video surveillance system are in a pilot stage being done independently and in isolation by local user departments at the city level. Diversity of technology solutions with varying levels of sophistication for addressing different traffic and road safety problems, diversity of suppliers of equipment (with many being startups), and non-standardisation make comparisons difficult. Chennai-based Geetha Krishnan, Professor in Civil Engineering, who has been involved in providing road safety solutions, agrees about the lack of study on the impact of surveillance and enforcement of traffic violations using emerging technologies.
Ravi Tiwari, Founder-CEO, BizAugmentor Global Services, a company involved in development of solutions for traffic management, is of the view that the different levels of sophistication, stages of implementation, and approaches found across cities in the country calls for some standardisation that would enable scalability and assessment of evidence backed enforcement. A committee may be formed under the aegis of the Ministry of Road Transport and Highways so that experts and user agencies may sit together and articulate real life road safety situations, evolve appropriate solutions and quality, says Ravi Tiwari.
The Road Transport Ministry seems to have taken note and is preparing a response to the situation. The government has identified effective enforcement as a focus area in its 13-point road safety policy toolkit and has announced its intent to assist various state and other governments to strengthen and improve the quality of enforcement in order to ensure effective and uniform implementation of safety laws. Piyush Jain, Director, MVL and IT, Ministry of Road Transport and Highways, said that guidelines were being prepared to help states implement the provision of the Motor Vehicles (Amendment) Act 2019.
This article is written as part of the World Health Organisationâ€™s Media Fellowship on Road Safety 2019.
G Rajasekaran is a senior journalist based in Salem and reports about development issues in the western districts of Tamil Nadu.