Application Study on Detection of Pipeline Weld Defects Based on SVMs

摘要

A new method which is based on Support Vector Machines (SVMs) for the identification of pipeline weld defects is proposed. In order to enhance the quality of the image, a lot of actions, such as image enhancement, morphological processing and edge detection, have been dealt with. As a result, many problems, such as excessive noise, fuzzy edge and low contrast, have been solved and it's beneficial to extract the features of the image. Firstly, the results of the identification of the second category are given. Then combined with the characteristics of multiple classifications, three structures of clustering are presented and the structure of one against one has been adopted to identify the samples after analysis. The experimental results show that the proposed model has a lot of advantages, such as the identification accuracy, high speed, easy to implement, etc, and it's suitable for identification of pipeline weld defects.