Algorithm Research Based On The Detection Of Moving Excess Objects

In recent years,with the rapid development of my country's aerospace industry,due to the high cost of aerospace products,the existence of poor reusability and other factors,the technology of detecting active waste has become more and more important.The production of excess activities is mainly caused by irregular production process,improper operation of workers and other reasons.The existence of redundant objects may cause the entire product to fail.An important indicator of aerospace products is reliability,so the detection of moving excess has become an urgent problem to be solved.In this context,this subject proposes an algorithm for detecting the material of the active redundant objects.First,with the help of a moving waste detection device based on the Particle Impact Noise Detection(PIND)method,the measured object is placed on the detection turntable,and the moving waste signal is detected by the PIND method.According to the detected signal of the moving waste,a suitable wavelet packet is selected for signal decomposition,the energy of each frequency band after the decomposition is calculated,and the moving waste signal is reconstructed according to the ratio,and the noise is filtered at the same time.Then the reconstructed signal is extracted in time domain and frequency domain.And for reference to the extraction method of Mel Frequency Cepstrum Coefficient(MFCC)eigenvalues in the speech signal,the MFCC eigenvalues are extracted from the active extras.Perform principal component analysis on the obtained eigenvalue samples,and determine the principal components according to the variance contribution rate to achieve the purpose of data dimensionality reduction.This process not only achieves the dimensionality reduction of the eigenvalue samples,but also retains the original eigenvalue samples The main information of the set.Finally,using the support vector machine method,the dimensionality-reduced feature value samples are divided into training samples and test samples.The training sample is used as the input parameter of the support vector machine classifier,and the classifier model generated by the training sample is used to predict the sample category of the test sample,so as to automatically detect the material of the active redundant object.Through experimental verification,the classification accuracy rate is as high as 90%.This paper adopts a research method that combines theory and practice,and refers to the existing algorithms for detecting active remnants to ensure that the algorithm for detecting active remnants meets the experimental requirements,and improves the accuracy of theoptimization algorithm and the quality of the active remnant detection materials.Accuracy.