基于群体智能算法的排球高鲁棒性目标识别研究（英文）
|
摘要
基于小波变换和人工神经网络的目标识别是图像处理的一个重要研究方向。但是,此类方法采用的梯度下降规则容易产生局部极小值。为了解决该问题,提出了一种基于群体智能算法的高鲁棒性目标识别算法,可有效应用于各种图像识别任务,如排球目标识别等。首先对图像进行预处理并变换成HSV空间进行背景分割,并通过小波不变矩对图像进行特征提取。然后采用新兴的群智能算法-狼群算法,对基于小波神经网络的目标图像识别进行优化,以便提升全局收敛性和鲁棒性。仿真实验结果显示:相比原有的方法,提出优化方法具有更高的识别精度和稳定性。
Target recognition based on wavelet transform and artificial neural network is an important research direction of image processing. However,gradient descent rules used by such methods tend to produce local minima.In order to solve this problem,a highly robust target recognition algorithm based on swarm intelligence algorithm was proposed,which can be effectively applied to various image recognition tasks such as volleyball target recognition. Firstly,the image was preprocessed and transformed into HSV space for background segmentation,and the features were extracted by wavelet moment invariants. Then using the new swarm intelligence algorithm-wolf pack algorithm,the target image recognition based on wavelet neural network was optimized to improve global convergence and robustness. Simulation results show that compared with the original method,the proposed optimization method has higher recognition accuracy and better stability.
Target recognition based on wavelet transform and artificial neural network is an important research direction of image processing. However,gradient descent rules used by such methods tend to produce local minima.In order to solve this problem,a highly robust target recognition algorithm based on swarm intelligence algorithm was proposed,which can be effectively applied to various image recognition tasks such as volleyball target recognition. Firstly,the image was preprocessed and transformed into HSV space for background segmentation,and the features were extracted by wavelet moment invariants. Then using the new swarm intelligence algorithm-wolf pack algorithm,the target image recognition based on wavelet neural network was optimized to improve global convergence and robustness. Simulation results show that compared with the original method,the proposed optimization method has higher recognition accuracy and better stability.
引文
[1]WANG K,QIU H,YANG H P.Network security situation assessment method based on attack pattern recognition[J].Computer application,2016,36(1):194-198.
[2]WANG Y K,GAO W X,TANG N,et al.Human posture recognition based on fuzzy pattern recognition[J].Computer engineering and design,2016,37(6):1621-1625.
[3]STACH S,BENARD J,GIURFA M.Local-feature assembling in visual pattern recognition and generalization in honeybees.[J].Nature,2017,42(3):758-761.
[4]FORREST S,JAVORNIK B,SMITH R E,et al.Using genetic algorithms to explore pattern recognition in the immune system[J].Evolutionary Computation,2014,1(3):191-211.
[5]LIN C M,TING A B,HSU C F,et al.Adaptive control for mimo uncertain nonlinear systems using recurrent wavelet neural network[J].International Journal of Neural Systems,2012,22(01):37-50.
[6]AVCI E,COTELI R.A new automatic target recognition system based on wavelet extreme learning machine[J].Expert Systems with Applications,2012,39(16):12340-12348.
[7]RUOHONG H,ZHANG P.SAR target recognition using PCA,ICA and Gabor wavelet decision fusion[J].Journal of Remote Sensing,2012,16(2):262-274.
[8]ISLAM M S,CHONG U P.Improvement in Moving Target Detection Based on Hough Transform and Wavelet[J].Iete Technical Review,2015,32(1):46-51.
[9]TUA S,EFE M.Continuous Wavelet Transform and Hidden Markov Model Based Target Detection[J].Radioengineering,2014,23(1):96-103.
[10]WU H S,ZHANG F M.Wolf Pack Algorithm for Unconstrained Global Optimization[J].Mathematical Problems in Engineering,2014(1):1-17.
[11]SHU L Y.Research about Underwater Target Recognition Based on Neural Network and Wavelet[J].Computer Simulation,2011,28(2):232-191.
[2]WANG Y K,GAO W X,TANG N,et al.Human posture recognition based on fuzzy pattern recognition[J].Computer engineering and design,2016,37(6):1621-1625.
[3]STACH S,BENARD J,GIURFA M.Local-feature assembling in visual pattern recognition and generalization in honeybees.[J].Nature,2017,42(3):758-761.
[4]FORREST S,JAVORNIK B,SMITH R E,et al.Using genetic algorithms to explore pattern recognition in the immune system[J].Evolutionary Computation,2014,1(3):191-211.
[5]LIN C M,TING A B,HSU C F,et al.Adaptive control for mimo uncertain nonlinear systems using recurrent wavelet neural network[J].International Journal of Neural Systems,2012,22(01):37-50.
[6]AVCI E,COTELI R.A new automatic target recognition system based on wavelet extreme learning machine[J].Expert Systems with Applications,2012,39(16):12340-12348.
[7]RUOHONG H,ZHANG P.SAR target recognition using PCA,ICA and Gabor wavelet decision fusion[J].Journal of Remote Sensing,2012,16(2):262-274.
[8]ISLAM M S,CHONG U P.Improvement in Moving Target Detection Based on Hough Transform and Wavelet[J].Iete Technical Review,2015,32(1):46-51.
[9]TUA S,EFE M.Continuous Wavelet Transform and Hidden Markov Model Based Target Detection[J].Radioengineering,2014,23(1):96-103.
[10]WU H S,ZHANG F M.Wolf Pack Algorithm for Unconstrained Global Optimization[J].Mathematical Problems in Engineering,2014(1):1-17.
[11]SHU L Y.Research about Underwater Target Recognition Based on Neural Network and Wavelet[J].Computer Simulation,2011,28(2):232-191.