Development of crowd detection warning system based on deep convolutional neural network using CCTV
Autor: | Muhammad Nurwidya Ardiansyah, Marifa Kurniasari, Muhammad Dzulfiqar Amien, Danang Wijaya, Pradana Setialana |
---|---|
Rok vydání: | 2022 |
Zdroj: | Journal of Engineering and Applied Technology. 3:35-41 |
ISSN: | 2716-2265 2716-2257 |
DOI: | 10.21831/jeatech.v3i2.43771 |
Popis: | The 2019 corona virus (Covid-19) pandemic is a global problem for now. One way to deal with the spread of the corona virus is to maintain a distance of at least one meter and stay away from crowds. Therefore, a crowd detection warning system based on a deep convolutional neural network (deep CNN) was developed using CCTV. The development of this system was carried out using the NVIDIA Jetson Nano microcontroller as the computing hardware. Crowd object detection uses the OpenCV library, the YOLOv3-Tiny algorithm, and the euclidean distance method to calculate the distance between 'person' objects. Based on the tests carried out on function and performance, the results obtained that this crowd detection warning system can detect 'person' objects with an accuracy rate of 92.79. In addition, this system has also been able to detect several types of colors from objects so that warning messages can be given more specifically on the color of the clothes of the 'person' in the detected crowd. |
Databáze: | OpenAIRE |
Externí odkaz: |