Abstract: Considering the disadvantages in the traditional image edge detection methods, such as the rough edge, noise of the edge and inaccurate edge location, in the study of plant roots, using the traditional image edge detection method to detect the edge cant obtain satisfactory result. A new efficient image edge detection algorithm method based on support vector machine (SVM) was proposed to solve above problems. Based on least squares SVM with Gaussian radial basis function kernel, a set of the new gradient operators and the corresponding second derivative operators are obtained. The performance of the presented edge detection algorithm is compared with Canny and Prewitt detectors. The experimental result demonstrated that, compared with conventional detection methods, the proposed edge detection could not only improve edge detection properties, but also could overcome the noise interference to a certain degree．

Key words: plant roots, edge detection, support vector machine, Gaussian radial basis function kernel