多光谱图像纹理特征数据挖掘方法仿真

为了更好地实现多光谱图像特征数据处理效果,将数据挖掘引入到多光谱图像特征数据处理中。但当前多光谱图像纹理特征数据挖掘过程中,普遍存在着特征数据挖掘时间过长、成本消耗过大、数据挖掘精确度较低等问题。提出基于Contourlet变换的图像纹理特征挖掘方法。对多光谱图像纹理特征数据进行模糊预处理,采用邻近范围相关性等知识去除多光谱图像包络线,在此基础上对多光谱图像纹理特征进行分析,利用形态学滤波算子去除多光谱图像中的噪声点。引用Contourlet变换方法将多光谱图像从空间域变换到频率域,提取了变换分解后的多光谱图像低频子带和高频子带的特征向量,完成多光谱图像纹理特征数据挖掘。实验结果表明,所提方法挖掘得到的数据均匀度较好、深浅度适中,挖掘精度高,且所提方法挖掘时间较短、成本消耗较低。

Simulation of Multi-Spectral Image Texture Feature Data Intelligent Mining Method

In order to achieve the processing effect of feature data in multispectral image,data mining is introduced into the multispectral image feature data processing.Currently,the feature data mining time is too long,the cost consumption is too high,and the data mining accuracy is low.Therefore,a method to mine image texture feature based on Contourlet transform was proposed.At first,the fuzzy preprocessing was performed on multi-spectral image texture feature data,and then the enveloping line of multi-spectral image was removed by knowledge of proximity correlation.On this basis,the multi-spectral image texture features were analyzed.Meanwhile,the morphological filter operator was used to remove the noise points in multi-spectral image.In addition,Contourlet transform method was used to transform the multi-spectral image from the spatial domain to the frequency domain.Finally,feature vectors of low-frequency sub band and high-frequency sub band of multi-spectral image after transformation and decomposition were extracted.Thus,the texture data mining of multi-spectral image was completed.Simulation results show that the proposed method has good uniformity of data and high mining accuracy.Meanwhile,this method has short mining time and low cost.