A Target Classification Optimization Recognition and Information Extracting Method of Laser Fuze Detection System Based on Fuzzy C-Means Incremental Update and Neural Network

With the rapid development of laser fuze technology, the ability of fuze to recognize target has become increasingly important. To effectively improve the recognition rate of laser fuze against interference signal and target signal, based on the two-channel symmetric laser scanning detection system, this paper proposes a classification method based on fuzzy c-means incremental update under the condition of small sample test data. The classification model is established using the fuzzy c-means algorithm, which uses the time-domain entropy, frequency-domain entropy and frequency domain exponential entropy of the reflected laser echo signal as the feature variables; The classification model is adaptively adjusted by adopting an improved fuzzy c-means incremental update algorithm to obtain the high accuracy classification results; Combining with the target’s signal features extracted by discrete binary wavelet transform, the classification results of real target signal are optimally recognized by wavelet neural network. The experimental results show that compared with the wavelet neural network recognition method, the target recognition method proposed in this paper based on fuzzy c-means incremental update and neural network can achieve higher target recognition rate under conditions of low signal-to-noise ratio and high threshold signal-to-noise ratio. Moreover, fuzzy c-means incremental update method reduces time-consuming of classification model, significantly improves the anti-interference ability, the results provide reference significance for future applications of neural network recognition method in laser fuze detection systems.