自适应PCNN和改进C-V结合的遥感图像变化检测

Remote sensing image change detection based on adaptive PCNN and improved C-V model

为获取保留图像信息较完好的差异图并得到更好的变化检测结果,提出一种基于自适应脉冲耦合神经网络(PCNN)和改进Chan-Vese(C-V)模型的非监督的不同时相遥感图像的变化检测算法。用差值法、比值法对两幅遥感图像进行差异图获取;用自适应PCNN图像融合算法对两幅差异图进行融合,获取保留图像信息较好的差异图;用基于改进C-V模型的分割算法对融合后的差异图进行分割,得到变化检测结果图。实验结果表明,该算法具有很好的变化检测效果,总检测精度较高。

To obtain difference image retaining image information much better ,and to gain better change detection results ,an unsupervised change detection algorithm in remote sensing images based on adaptive pulse coupled neural network （PCNN） and improved Chan‐Vese model was proposed .Firstly ,the difference images were generated by using the difference method and ratio method on two remote sensing images .Then ,the results of the subtraction method and the ratio method were fused by using the adaptive PCNN algorithm to obtain complementary information .The change regions were separated from the merged difference image by using the image segmentation algorithm based on improved C‐V model to gain change detection results figure .The ex‐perimental results show that the algorithm has good effect on change detection and it has higher detection precision .