Forest fireworks identification intelligent cloud system based on distributed computing

Abstract

The invention discloses a forest firework identification intelligent cloud system based on distributed computing. The system includes a front end, a cloud, and a mobile terminal. The front end includes a front-end network communication module, a front-end image acquisition module, a front-end image acquisition auxiliary module, and a front-end power supply module. The cloud includes a network communication module, a central processing module, a parallel computing module storage module, and an alarm module. A number of front ends are set up at various monitoring points in the forest area to achieve feature extraction and transformation of the acquired real-time image. A cloud connects to multiple front ends through a forest area LAN or mobile Internet, obtains data and sets up a deep neuralnetwork for identification. The feature distribution of the video information with a large amount of data is extracted and transformed at the front end. The small amount of information after the feature transformation can be transmitted in real time without being affected by the low-speed data link. The amount of cloud computing is greatly reduced, and the recognition rate is also achieved and greatly improved in a complex environment.

Description

technical field [0001] The invention relates to an intelligent fire protection system, in particular to an intelligent cloud system for forest firework identification based on distributed computing. Background technique [0002] As an environmental ecological resource, forest resources play an important role in forestry and even environmental protection. Our country is vast in land and resources, and its forest area occupies a decisive position in the world. In recent years, the objective existence of unfavorable meteorological conditions, the complex social environment in forest areas and the imperfect forest fire prevention mechanism have made the form of forest fire prevention very severe. The property losses caused by forest fires are hundreds of millions, and the investment of various countries in forest fire prevention Very huge. [0003] my country's forest fire prevention construction is still at a relatively low level. In recent years, the investment in forest fir...

AI Technical Summary

Problems solved by technology

[0010] The use of computer deep learning can well solve the impact of environmental influences on pyrotechnic image recognition, but if the feature extraction of images is directly performed using deep neural networks, a large amount of bandwidth is required for video transmission and a huge amount of calculation, which affects transmission resources and calculations. The demand for resources is huge, and it is difficult for general servers to provide support

Images