4 days ago
Ahmedabad police to use AI-based system during Rath Yatra to avoid stampede. How does it work?
Ahead of the 148th Jagannath Rath Yatra on June 27, the Ahmedabad city police are setting up advanced surveillance systems to ensure the safety of around 15 lakh devotees who are expected to gather to watch the holy procession go through some of the most congested areas of the Old City area.
This becomes significant in light of the recent stampede in Karnataka's Bengaluru during the celebrations of the Royal Challengers Bengaluru's IPL victory that led to the death of 11 people and injured dozens of others. Notably, a stampede also took place at the Mahakumbh held in Uttar Pradesh's Prayagraj earlier this year.
While The Indian Express had earlier reported on newly launched GP-DRASTI drone programme that will, also for the first time, be deployed in a major 'moving' crowd event during the Rath Yatra later in June, now, another software, to be integrated into the CCTV cameras along the route as well as the drone programme, enabled with Artificial Intelligence (AI), will be used for 'stampede avoidance'.
This system, still in the development stage, and yet to be named, will calculate the density of people in a particular area of spanning certain square metres and determine, based on yet-to-be-decided parameters, if a stampede is imminent or not. Based on the results of this system, alerts will be sent out to officers to disperse crowds to avoid a stampede.
According to information shared by the Detection of Crime Branch (DCB), 'Anti-stampede algorithms on CCTV cameras are a crucial advancement in crowd management, leveraging AI and image processing to prevent dangerous situations in densely populated areas.'
ACP Bharat Patel from the Crime Branch said, 'Last year, we had used a lot of technology to monitor and handle the Rath Yatra, which starts from Jagannath Temple, held till Saraspur and ends at the same temple. This time, we are planning to use an AI-based stampede avoidance system especially at the most congested spots along the route of the yatra.'
The most congested spots along the Rath Yatra route are the Jagannath Temple, the narrow route near Dhal Ni Pol, the area near Chakleshwar Mahadev temple, the Saraspur stop where the participants stop for lunch, and then a narrow route in Shahpur area on the way back to the Jagannath Temple.
Speaking on the technology, ACP Patel said that the system will use thermal imaging and pixel-counting to count the occupied and unoccupied areas of a particular space and constantly update this information on the screens that will be placed in the police control room as well as field units so that ground forces can be immediately advised to take action if the number of people in a particular space exceed the threshold limit.
However, when asked about the threshold limit or density of people per square metre, the officer said that was yet to be decided, adding that it will vary based on areas and the varying egress points available.
Further, since the Rath Yatra is a procession or a moving event, the system will be able to calculate, based on crowd movement, if a large group of people is moving to an area, which may not be able to accommodate them. In such cases, police personnel will be asked to move along with the people already in that particular area, to make room for the incoming crowd.
While the officer stressed that there had not been any stampede at the yatra held in Ahmedabad in the recent years, the advanced technology was being used as a precautionary measure, taking lessons from incidents elsewhere.
1. Real-time monitoring: AI-powered CCTV cameras will continuously analyse video streams in real time.
2. Crowd density estimation: Algorithms will calculate the number of people in a given area. This can involve pixel-based analysis (converting images to black and white and counting 'black pixels' representing people), and object detection, using machine learning models to identify and count individuals by detecting heads or torsos.
3. Thresholding: Pre-defined 'threshold values' for crowd density will be established. When the detected density crosses these thresholds, it will trigger an alert.
4. Anomaly detection: Beyond just density, these algorithms can identify unusual crowd behaviour, such as sudden surges in movement, unusual clustering patterns, fallen individuals, and aggressive movement.
5. Alerting authorities: Upon detecting a potential stampede risk, the system sends immediate alerts to security personnel or control rooms via LCD displays, GSM messages or other communication channels.
6. Predictive analytics: Advanced systems will use time-series prediction models to forecast crowd behavior and dynamics based on historical and real-time data, helping anticipate potential bottlenecks or overcrowding.
7. Reinforcement learning: Algorithms can learn from past incidents to suggest optimal crowd flow routes and alternative evacuation paths during emergencies.
—–
1. Proactive prevention: The primary benefit is the ability to detect and warn of potential stampedes before they occur, allowing authorities to take preventative measures.
2. Real-time insights: Provides immediate and accurate data on crowd density and movement, far surpassing manual observation.
3. Enhanced safety: Significantly improves safety in public spaces by reducing human error and enabling swift responses to risks.
4. Optimised resource allocation: Helps in better deployment of security personnel and resources to areas with high crowd density.
5. Improved efficiency: Automates a labor-intensive task, freeing up human operators for more complex decision-making.
6. Data for future planning: The collected data can be analyzed to improve crowd management strategies for future events.
—–
1. Limited accuracy: AI algorithms can face challenges with occlusion (people blocking each other), varying conditions (changes in lighting, weather, and camera angles), and bias in training data (leading to false positives).
2. Computational complexity and cost: Developing and deploying such systems can be expensive due to the need for high-resolution cameras, powerful processing units, and sophisticated algorithms.
3. Data privacy and ethical concerns: The extensive use of CCTV and AI raises concerns about individual privacy and potential misuse of data.
4. Integration with existing infrastructure: Integrating new AI-powered systems with older CCTV networks can be complex.
5. Human intervention remains crucial: While AI can alert, human responders are still essential for effective intervention and crowd dispersal. As seen during Maha Kumbh, even with AI alerts, a lack of ground personnel can limit effectiveness.
6. Defining thresholds: Determining appropriate crowd density thresholds for different environments and cultural contexts can be challenging.
