A Character Superposition Method Based on Object Detection

Autor:	Jie Zhao, Li Hushan, Hu Lianyue, Cen Jian, Xu Wenkai
Rok vydání:	2020
Předmět:	Computer science Information value business.industry Monitoring system Object detection Constant false alarm rate 03 medical and health sciences 0302 clinical medicine Character (mathematics) Computer vision 030212 general & internal medicine Artificial intelligence Superposition method business
Zdroj:	Advances in Brain Inspired Cognitive Systems ISBN: 9783030394301 BICS
DOI:	10.1007/978-3-030-39431-8_31
Popis:	In this paper, the video data collected by the monitoring system in the urban underground integrated pipe gallery and the sensor data are independent of each other, which leads to the problem that the information value cannot be fully utilized. A character superposition method based on object detection is proposed. The method uses the sensor data as a character in the video, which solves the problem of information island between sensor data and video data. It is verified by experiments that the accuracy of object detection algorithm proposed by this paper is 13.94% higher than that of traditional target detection algorithm, and the false alarm rate is reduced by 14.29%. It solves the problem of accurate monitoring under the complex environment of urban underground integrated pipe gallery.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::82b80bfab1a5af03e9e1f6bbc5963df1 https://doi.org/10.1007/978-3-030-39431-8_31 Zobrazit plný text záznamu