基于输电线路悬挂异物航拍图像的图像分割算法研究

An Image Segmentation Algorithm Based on Power Line Aerial Images with Foreign Matters

针对复杂背景下输电线路悬挂异物航拍图像前景与背景的分割问题,提出一种可适应大量此类图像的图像分割方法。该方法由图像预处理、图像分割和图像分析三个步骤构成。图像预处理步骤主要完成航拍图像放大裁剪、滤波等工作;图像分割步骤提出并选用一种基于粒子群优化算法改进的OTSU双阈值分割算法对图像进行前景与背景的初分割,并选用SVM算法对图像进行再分割,据此大致完成将电力线和异物与背景分离的工作;图像分析步骤提出一种基于区域特征的二值图像处理方法,并据此完成将电力线和异物与背景分离的最终工作。最后,通过样本图片分析了算法的准确率和运算时间。结果表明,该算法能快速准确地对输电线路悬挂异物航拍图像进行前景与背景的分割。

For the image segmentation of power line aerial images with foreign matters,an image segmentation algorithm is proposed.This method is consisted of three steps,namely image pre-processing,image segmentation and image analysis.Firstly,the image pre-processing step mainly completes the amplification,clipping and filtering work of aerial images.Secondly,in the im⁃age segmentation step,the OTSU dual threshold algorithm based on particle swarm optimization algorithm is proposed and selected to perform initial segmentation of the foreground and background of the image,and SVM algorithm is selected to perform image re-segmentation.This step roughly completes the work of separating the power lines and foreign matters from the background.Third⁃ly,in the image analysis step,a binary image processing method based on region features is proposed to perform final segmentation.Finally,the accuracy rate and operation time in the whole algorithm process are analyzed with sample data.The result shows that this method can quickly and accurately segment the foreground and background of power line aerial images with foreign matters.