基于偏微分方程的遥感成像雷达距离图像分类方法

针对传统方法分类遥感成像雷达距离图像时, 未对图像进行平滑处理, 导致其易受环境干扰, 分类性能较差的问题, 提出一种基于偏微分方程的遥感成像雷达距离图像分类方法. 首先通过偏微分方程对遥感成像雷达距离图像进行平滑处理, 然后采用基于偏微分方程的多区域分割模型, 将分割后的遥感成像雷达距离图像分类过程视为泛化函数最小化过程, 通过分割对能量泛函数进行最小化处理, 实现遥感成像雷达距离图像的多区域分类. 实验结果表明, 该方法成像速度快, 去噪和图像分割效果好, 分类精度和Kappa系数值均较高.

Distance Image Classification Method for Remote Sensing Imaging Radar Based on PDE

Aiming at the problem that the traditional method of classification of remote sensing imaging radar range images did not smooth the image, which made it vulnerable to environmental interferenceand poor classification performance, the author proposed a distance image classification method for remote sensing imaging radar based on partialdifferential equation. Firstly, the author smoothed the distance image of remote sensing imaging radar by improving the partial differential equation. Secondly,a multi region segmentation model based on partial differential equation was used to treat the distance image classification process of remote sensing imaging radar as minimization process of generalization function, the multi region classification of the distance image of remote sensing imaging radar was realized by minimizing the energy function through the segmentation. Experimental results show that the proposed method has fast imaging speed,good denoising and image segmentation effect, high classification accuracy and Kappa coefficient value.