Design and Exploration of a Low-power Forest Fire Prevention Video Terminal Based on Edge Computing
In response to the difficulties of power and network supply in forest areas in Beijing,the large storage space occupied by forest fire video surveillance data,and the low rate of manual identification,a low-power forest fire video terminal box has been designed using Deep Convolutional Neural Networks.It features real-time monitoring,automatic identification,and on-demand back transmission capabilities.This device can automatically monitor,identify,alert,and back transmit images of suspected fires at the forest fire video terminal,automatically filtering out over 90%of repetitive and invalid video images,saving storage resources and reducing manual labor costs.