摘要
针对传统的基本测地线活动轮廓(GAC)模型在检测噪声干扰、弱边界及凹陷边界目标的轮廓时提取效果不佳的问题,文中提出一种基于灰色关联分析的改进GAC模型轮廓检测方法。该方法利用灰色关联度代替梯度信息来构建停止函数。与传统的梯度信息相比,灰色关联系数对于具有模糊的边界信息的图像信息表示更为准确,从而更好地提取弱边界目标轮廓。初步实验结果表明,文中方法在提取弱边界目标轮廓时效果优于基于传统GAC模型和传统的LBF模型的轮廓检测方法。
In view of the problem of bad extraction effect when traditional basic GAC model detects the noise and contours of object con-taining concave edges or weak edges,propose an improved contour detection method of GAC based on grey relational analysis in this pa-per. In this method,use the grey relational coefficients instead of gradient information to construct the stop function. Compared with tradi-tional gradient information,a grey relational coefficient is more accurate in expressing the image information with fuzzy boundary,which extracts object contours with weak edges well. Preliminary experimental results show that the presented method is better than the contours detection method based on traditional GAC model and LBF model in detecting weak edges.
出处
《计算机技术与发展》
2015年第1期70-73,共4页
Computer Technology and Development
基金
国家自然科学基金资助项目(61202153)
陕西省语音与图像处理重点实验室开放基金项目(SJ2013002)
