Person Detection for Social Distancing and Safety Violation Alert based on Segmented ROI
In addressing the worldwide Covid-19 pandemic situation, the process of flattening the curve for coronavirus cases will be difficult if the citizens do not take action to prevent the spread of the virus. One of the most important practices in these outbreaks is to ensure a safe distance between peop...
| 总结: | In addressing the worldwide Covid-19 pandemic situation, the process of flattening the curve for coronavirus cases will be difficult if the citizens do not take action to prevent the spread of the virus. One of the most important practices in these outbreaks is to ensure a safe distance between people in public. This paper presents the detection of people with social distance monitoring as a precautionary measure in reducing physical contact between people. This study focuses on detecting people in areas of interest using the MobileNet Single Shot Multibox Detector (SSD) object tracking model and OpenCV library for image processing. The distance will be computed between the persons detected in the captured footage and then compared to a fixed pixels' values. The distance is measured between the central points and the overlapping boundary between persons in the segmented tracking area. With the detection of unsafe distances between people, alerts or warnings can be issued to keep the distance safe. In addition to social distance measure, another key feature of the system is detecting the presence of people in restricted areas, which can also be used to trigger warnings. Some analysis has been performed to test the effectiveness of the program for both purposes. From the results obtained, the distance tracking system achieved between 56.5% to 68% accuracy for testing performed on outdoor and challenging input videos, while 100% accuracy was achieved for the controlled environment on indoor testing. Whereas for the safety violation alert feature based on segmented ROI, it was found to have achieved better accuracy, i.e. between 95.8% to 100% for all tested input videos. © 2020 IEEE. |
|---|---|
| ISSN: | |
| DOI: | 10.1109/ICCSCE50387.2020.9204934 |