由于小波分解的多分辨分析特征及神经网络的自学习、自组织等性能,在图像处理中得到了广泛的应用。本文研究了SAR图像非线性采样目标低频小波树特征提取方法,利用PCA(主分量分析)对低频小波树降维,用降维后的特征值训练LVQ神经网络,将其应用于SAR图像目标检测,取得了较好的检测结果。%Multi-resolution analysis (MRA) characteristic of wavelet transform and self-learning and self-organiza- tion capability of neural network have been widely used in image processing. A target feature extraction method based on low frequency wavelet tree in nonlinear sampling of SAR image is studied. Dimensions of low frequency wavelet tree is reduced by way of principal component analysis (PCA), results obtained are used to train learning vector quantization (LVQ) neural network; good detection results can be obtained when the trained neural network is applied in target detection in SAR image.
展开▼