基于多特征的成捆原木端面轮廓识别方法

摘要

为了准确地识别原木端面轮廓，建立了联合像素特征和边缘结构的成捆原木轮廓识别模型，引入支持向量机改进原木端面像素点的识别效果，并用最小二乘法对边缘轮廓进行圆拟合。实验结果表明，在图像背景、原木端面都不理想的情况下，该方法能较为有效地完成成捆原木端面轮廓识别。%A model of outline eztraction of bundled logs cross section based upon pizel feature and edge structure is developed to improve bundled logs image processing. A support vector machine(SVM)is used to eztract the pizels of cross section of logs,and the least square method is applied to fit the circle of edge pizels. Finally,outlines of bundled logs are eztracted. The obtained results of ezperiment demonstrate the validity of the proposed method in the case of the cross section of logs is not the ideal smooth and the background is complez.