A new family monitoring alarm system based on improved YOLO network
Chengtao Cai, Boyu Wang, Xin Rong Liang · 2018
The traditional home remote monitoring has many problems. For example, the system can not independently identify the invading species, the system can not have targeted automatic alarm, and the accuracy of the system identification target is greatly affected by the external environment and other issues. About these problems, this paper presents a method based on deep learning, which can accurately find and distinguish the types of invaders by using the machine to train the characteristics of multiple species. This system improves the parameters of YOLO Network, it improves the processing speed, it also complete the real-time alarm, local storage and real-time remote playback. The function of the system is not affected by the external environment. In a complex environment for multiple experiments, the experimental results show that the intelligent monitoring system based on the improved YOLO network has high detection precision and fastness, and the false detection rate is 0 in the background of home remote monitoring. The program is feasible.