支持向量机理论在植物根系图像边缘检测中的应用  

Application of Support Vector Machine Theory to Image Edge Detection of Plant Roots

作　　者：吴鹏[1] 宋文龙[1] 

机构地区：[1]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2012年第8期153-156,共4页Journal of Northeast Forestry University

基　　金：中央高校基本科研业务费专项资金(Z02068)

摘　　要：由于传统边缘检测方法中存在的比如粗糙边缘、噪声边缘和不准确边缘等缺点,因此在植物根系的研究中,采用传统的图像边缘检测方法检测出来的边缘信息都无法达到令人满意的效果。本文基于带高斯径向基核函数的最小二乘支持向量机方法,得到了一簇梯度算子和零交叉算子,用来定位图像边缘,从而得到一种有效的图像边缘检测算法。用所得到的边缘检测算法与Sobel算法和Prewitt算法的性能进行了比较。仿真结果表明本文给出的算法与传统算法相比,不仅边缘检测性能得到提高,而且可以一定程度地克服噪声干扰。Traditional image edge detection methods of plant roots exhibit some disadvantages, such as rough edge, marginal noise and inaccurate edge location. The edge detection of plant roots by the traditional image edge detection method can＇ t obtain satisfactory results. Based on the least squares support vector machine （SVM） with radially Gaussian kernel function, a set of new gradient operators and zero crossing operators are obtained. An efficient image edge detection algorithm, using the gradients to locate the edge position, is proposed to solve the above problems. The performance of the presented edge detection algorithm is compared with those of Canny and Prewitt detectors. Compared with the conventional detection methods, the detection performance of the proposed edge detection method is improved, and the noise interference, to a certain extent, can also be overcome.

关 键 词：植物根系 边缘检测 最小二乘支持向量机 高斯径向基核函数 零交叉算子 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]