基于自适应权重法的K-means模型对遥感图像分割

针对传统K-means算法不易获得最优质心及易于趋向局部最优的问题,提出一种基于最优权重法的K-means模型对遥感图像分割的方法。使用二维高斯函数对遥感图像进行滤波平滑处理,减少噪声对像素点的影响;依据早熟收敛度和自适应值进行调整,找到最优权重作为初始聚类中心,从而有效地跳出局部最优;将样本分配到每个聚类中心,不断进行迭代更新簇中心,直至算法最终收敛。实验结果表明,该算法的分割精度有较明显的提高。与传统的K-means分割算法及GA分割算法相比,该算法对遥感图像分割的效果更为明显。

K-MEANS MODEL BASED ON ADAPTIVE WEIGHT METHOD FOR REMOTE SENSING IMAGE SEGMENTATION

Aiming at the problem that the traditional K-means algorithm is not easy to obtain the highest quality and tends to be local optimal,we proposed K-means model based on optimal weight method for remote sensing image segmentation.We used two-dimensional Gaussian function to filter and smooth the remote sensing image to reduce the influence of noise on the pixel.It adjusted according to the premature convergence and adaptive value to find the optimal weight as the initial clustering center,and effectively jumped out the local maximum.Samples were allocated to each cluster center,and the cluster center was updated iteratively until the algorithm finally converged.The experimental results show that the segmentation accuracy of the algorithm is improved obviously.Compared with the traditional Kmeans segmentation algorithm and GA segmentation algorithm,the proposed algorithm is more effective for remote sensing image segmentation.