http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
적외선 영상 기반 신경망 구조와 군집 분석을 이용한 고속 피플 카운팅
권현송 (Hyun-Song Kwon),이종화(Jong-Whoa Lee),구호근(Ho-Geun Koo),이범주(Buem-Joo Lee),김영국(Young-Kuk Kim) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
"People Counting" is one of the main research areas of object detection and counting algorithms. This algorithm uses various electronic sensors or image cameras to automatically count the number of people passing by. In this paper, high—speed people counting algorithm using neural network structure in infrared video is proposed. The advantage of this method is that it is fast and can also be trained with classification data which is easy to build dataset. The proposed algorithm is performed in the order of detecting pedestrians using neural network and cluster analysis. After that it analyzes the movements of the detected pedestrians and counting the number of in—out people. Furthermore, the corresponding algorithm showed 3.03 times faster processing speed than the YOLOv3—tiny model which is specialized in high—speed operation of object detection.