Research Of Automatic Identification Technology In Microwave Fluctuation Impending Earthquake

The phenomenon of microwave fluctuation impending strong earthquake is a kind of abnormal pulsation record which dominant frequency of spectrum looks pretty orderly and focuses on 11-16 Hz.It was reproduced before Wenchuan,Yushu,Qingchuan and other strong earthquakes,and it can be used for judgment of impending strong earthquake.At present,the artificial method for the identification of the phenomenon of microwave fluctuation impending earthquake was used to the real-time monitoring technology system of ground pulse in Lanzhou(surveillance system of Ground Pulse for short),which mainly depends on the experience of the technicians,and the subjective factors have a great influence on the recognition accuracy.By comparing the similarity between the waveform to be detected and the sample waveform,image sensing hashing technology was realized for the automatic recognition of microwave fluctuation impending earthquake.The article summarizes the relevant theories of perceptual hashing and its basic framework for image recognition.Meanwhile,combined with the frequency domain characteristics of the microwave fluctuation impending earthquake,two methods of automatic sensing based on block mean and DCT transform are proposed,two major earthquake events which occurred in 2008 of Wenchuan Ms8.0 and in 2010 of Yushu Ms7.1 as typical earthquake cases,by using the continuous pulsation records of the remote station within 250 KM and some remote stations,the operating efficiency and accuracy of the two sensing algorithms is tested by 16*16,32*32,64*64 three image compression sizes,in order to select the "best" algorithm for automatic identification of microwave fluctuation impending earthquake.The experimental results show that,in terms of recognition accuracy,the recognition accuracy of DCT transform is higher than that of block mean,and different image compression sizes have great influence on the two algorithms.In terms of operation efficiency,image compression size has a great influence on DCT transform algorithm,and block mean algorithm is almost unaffected by it.the DCT transform algorithm with image compression size of 32*32 has high accuracy and operational efficiency for the identification of the microwave fluctuation impending earthquake,and the correlation coefficient between Wenchuan earthquake and Yushu earthquake are 0.81 and 0.71 respectively;the algorithm operating efficiency is 1175 suspected events and 2057 suspected events per minute,respectively.The results of two earthquake cases test show that only stations within the range of 50 KM can automatically identify the anomalies of microwave fluctuation impending earthquake within 21 days before the earthquake,and there are no anomalies beyond 50 KM or 21 days before the earthquake,which accords with the tracking index based on Microwave fluctuation impending earthquake.It is proved that the algorithm can be used for automatic identification of Microwave fluctuation impending earthquake and realizing earthquake tracking.