Building Health State Recognition Method Based on Multi-channel Convolution Neural Network Fusion

The identification of the health status of buildings has been paid more and more attention by all sectors of the society. The early warning of catastrophes or the assessment of the damage degree and residual life of building structures after catastrophes has become a hot topic for scholars from all over the world. In order to improve the performance of building health state recognition, a novel framework based on multi-channel convolution neural network fusion is proposed in this paper. By combining the output results of different convolution neural networks, temporal information and spatial information are used to achieve the accurate classification of building health status. Eventually, with the data collected by the sensor during the earthquake, the proposed framework is proved to be effective and superior.