Physical distancing, defining a “new normal” as a measure to contain the spread of Covid-19. Unless they are members of a family, pedestrians in shared spaces are required to maintain a minimum (country-dependant) pairwise distance. Managers of public spaces may be tasked with enforcing or monitoring this constraint. As privacy-respecting real-time tracking of pedestrian dynamics in public spaces becomes more common, it is natural to use these tools to analyse the adherence to physical distancing and compare the effectiveness of crowd management measurements.
Crowd detection analytics can be used in any environment to limit occupancy and detect overcrowding. Because there are no exceptions for social distancing, crowd detection could be a valuable tool in ensuring that physical distancing is observed in all settings.
To address this, several cross-faculty students at Indonesia’s Universitas Gadjah Mada (UGM) developed a crowd detection system to prevent COVID-19 transmission. The system was said to be built using Deep Learning and WebGIS. As a result, the tool can detect crowds by providing data on the number of people in the crowd. and displays a visualisation of field conditions, including the time and location where the crowd occurs in “near real-time.”
The crowd detection prototype was invented in 2021 as part of the Student Creativity Programme in the Field of Copyright (PKM-KC), which received Rp. 9 million development grants from the Ministry of Education and Culture. The system also has an early warning feature for crowds. A crowd warning will be transmitted automatically via loudspeakers at the detected location. He explained that the device detects crowds using visual data input obtained from CCTV via a webcam connected to a local computer that has previously been programmed with deep learning to detect human presence and predict crowded situations.
Once done, the data results are sent to the web technology and the Geographical Information System (WebGIS) in the form of information related to the location, time and the number of crowd events in one location monitored by CCTV. to give warnings,” said the researcher. Later, they plan to add a text alert feature to make it easier for officers to monitor.
Per another researcher, the development of crowd detection began with concerns that there were still many violations of health protocols occurring in the community, particularly regarding social distancing and avoiding crowds. “Keep your distance and avoid crowds because the monitoring of the apparatus is not optimal. Therefore, we took the initiative to develop this detection tool to make it easier for officers to monitor and take immediate action,” he said.
The system, which has been in development since June 2021, has been field-tested. As a result, it detects crowds in a room with greater than 75% accuracy. “Even with a webcam, it can produce quite a good accuracy for detecting crowds with medium and low image resolutions. However, in the future, high-resolution CCTV will be developed so that results could be more accurate,” he said.
OpenGov Asia reported that throughout this pandemic, the healthcare sector has been in urgent need of healthcare management systems to manage hospitals, clinics and other medical facilities; the demand for healthcare tech has been soaring. Artificial intelligence is progressively being seen as an excellent technology to leverage for healthcare, even as it becomes more prevalent in modern business and everyday life.
The significance of this information can aid in the release of venues from lockdowns, which would necessitate a high level of trust from those most affected (e.g., front-line staff such as teachers, waitresses, shop workers, etc.). Putting crowd detection tools in place shows these key stakeholders that the venue is doing everything possible to keep people safe.
