Split Bregman算法在遥感图像边缘检测中的应用研究

发布时间：2018-04-30 06:35

本文选题：边缘检测 + 主动轮廓模型　；参考：《计算机科学》2017年12期

【摘要】：针对基于水平集的边缘检测方法抗噪性能差,处理模糊边界和灰度不均匀性的能力弱,计算效率低,边缘检测结果的精确性极大程度上依赖于初始轮廓的位置且曲线演化易陷入极小值等问题,提出一种基于全局最优凸函数变分模型和Split Bregman数字最小化的边缘检测方法。该方法首先根据CV模型的构造原理,并利用Chan的全局最优化思想,构造了一个通用的可以获得全局最优解的凸函数变分模型;在曲线演化和数字最小化迭代过程中,引入了基于Split Bregman的快速迭代算法,可以使曲线在克服噪声等问题的影响下快速、准确、稳定地演化到目标的边界处。实验结果证明了提出的边缘检测方法既具有较高的计算效率,能够满足遥感图像边缘检测对实时性的要求,同时也具有较高的边缘检测精度和良好的普适性。
[Abstract]:The edge detection method based on level set has poor anti-noise performance, weak ability to deal with the inhomogeneity of fuzzy boundary and gray scale, and low computational efficiency. The accuracy of the edge detection results depends on the position of the initial contour to a great extent and the curve evolution is easy to fall into the minimum. A new edge detection method based on the global optimal convex function variational model and the minimization of Split Bregman numbers is proposed. Based on the construction principle of CV model and Chan's idea of global optimization, this method constructs a general variational model of convex function, which can obtain the global optimal solution, in the process of curve evolution and digital minimization iteration. A fast iterative algorithm based on Split Bregman is introduced, which can make the curve evolve to the boundary of the target quickly, accurately and stably under the influence of noise and other problems. The experimental results show that the proposed edge detection method not only has a high computational efficiency, can meet the real-time requirements of remote sensing image edge detection, but also has a higher edge detection accuracy and good universality.
【作者单位】：大连外国语大学软件学院;
【基金】：国家自然科学基金项目(61501082,61502435) 辽宁省自然科学基金项目资助(L2015020017,L20170540232,L20170540230) 辽宁省教育厅科学研究一般项目(L2015137,L2014455) 辽宁省社会科学规划基金项目(L15CGL009)资助
【分类号】：TP751

【相似文献】