Stress reaction process-based hierarchical recognition algorithm for continuous intrusion events in optical fiber prewarning system

Pattern Recognition of Grating Perimeter Intrusion Behavior in Deep Learning Method

“…With reference to the above experimental setup, according to this method, the characteristic of fiber vibration signal was extracted, obtaining the characteristics of intrusion signals such as shaking protection nets, knocking protection nets, and over-protection fences, and obtaining their respective feature sets after condensing. The method of reference from [4], the method of reference from [5], and the proposed method were used to identify and identify the intrusion signals. The recognition results were obtained by identifying and comparing the features.…”

“…Dividing the correct recognition of each category by the number of test signals and calculating the accuracy of their identification, we received the results shown in Table 1. According to the data in Table 1, the average recognition accuracy rate of intrusion signal of reference [4] method is 79.9%, that of reference [5] method is 76.6%, and that of the proposed method is as high as 94.4%. It can be seen that the recognition accuracy of intrusion signals such as the shaking protection net, striking protection net, and crossing the protection fence of the proposed method are higher than those of the reference [4] method and the reference [5] method.…”

“…According to the data in Table 1, the average recognition accuracy rate of intrusion signal of reference [4] method is 79.9%, that of reference [5] method is 76.6%, and that of the proposed method is as high as 94.4%. It can be seen that the recognition accuracy of intrusion signals such as the shaking protection net, striking protection net, and crossing the protection fence of the proposed method are higher than those of the reference [4] method and the reference [5] method. The recognition accuracy rate of intrusion signals includes the net and over the fence.…”

“…This method can realize more accurate and effective intrusion recognition. In [5], a hierarchical recognition algorithm for continuous intrusion events based on the stress response process was proposed. Using the continuity of intrusion events, a heuristic staged recognition process was established, and the stress response degree of intrusion events was analyzed.…”

“…In this part, the amplitude of the intrusion vibration signal was measured by the experimental environment, and the noise problem was also considered. The method in this paper was compared with the method in [4,5], and the accuracy of intrusion signal recognition, the iteration loss value of training set, the identification efficiency of intrusion events, and the response sensitivity of intrusion events were compared and analyzed. The results showed that the accuracy and efficiency of this method were high, the false alarm rate of the intrusion signal was greatly reduced, and the overall stability of the recognition system was improved.…”