DT-CWT结合MRF的遥感图像变化检测

发布时间：2018-06-08 15:38

本文选题：变化检测 + 双树复小波变换DT-CWT　；参考：《遥感学报》2017年03期

【摘要】：为了解决多尺度遥感图像变化检测在降噪时丢失大量高频信息及单一像素孤立性的问题,提出了一种双树复小波变换DT-CWT(Dual-tree Complex Wavelet Transform)和马尔可夫随机场MRF(Markov Random Field)相结合的非监督遥感图像变化检测算法,首先采用DT-CWT对差异图像进行多尺度分解,并根据MRF模型分割算法提取高频区域的变化特征,然后进行相应层的高、低频重构,再对重构后的各层建立MRF模型并根据贝叶斯最大后验概率准则MAP(Maximum A Posterior)进行最终分割,最后对各层分割结果进行求交融合,得到最终的变化检测结果掩膜图。对比实验结果表明,该方法在去除杂点和噪声的同时能够较好地保留高频信息,并且边缘检测更加平滑,具有较高的变化检测精度和很好的鲁棒性。
[Abstract]:In order to solve the problem of multi-scale remote sensing image change detection losing a lot of high-frequency information and single pixel isolation in noise reduction. An unsupervised remote sensing image change detection algorithm based on DT-CWT-Dual-tree complex Wavelet transform (DT-CWT-Dual-tree complex Wavelet transform) and Markov Random Field (MRFF) is proposed. Firstly, DT-CWT is used to decompose the differential image. According to the MRF model segmentation algorithm, the change characteristics of the high frequency region are extracted, and then the high and low frequency reconstruction of the corresponding layer is carried out. Then, the MRF model is established for each reconstructed layer and the final segmentation is carried out according to the Bayesian maximum posterior probability criterion (MAPMaximum A Posterior). Finally, the intersection fusion of each layer segmentation result is carried out, and the final change detection result mask image is obtained. The experimental results show that the proposed method can preserve the high frequency information while removing the clutter and noise, and the edge detection is smoother, with high accuracy and robustness.
【作者单位】：中国矿业大学环境与测绘学院;国家测绘地理信息局卫星测绘应用中心;
【基金】：国家自然科学基金(编号:51374209,41271394)~~
【分类号】：TP751

【相似文献】