基于NSCT的航拍绝缘子图像边缘提取方法  

Aerial insulator image edge extraction method based on NSCT

作　　者：赵振兵[1] 金思新[1,2] 刘亚春[1] 

机构地区：[1]华北电力大学电气与电子工程学院,保定071003 [2]华东交通大学信息工程学院,南昌330013

出　　处：《仪器仪表学报》2012年第9期2045-2052,共8页Chinese Journal of Scientific Instrument

基　　金：中央高校基本科研业务费专项资金(12MS122)资助项目

摘　　要：绝缘子图像边缘提取是实现航拍绝缘子缺陷检测与识别的重要前提,结合航拍绝缘子图像的特点,提出了一种基于非下采样轮廓波变换(non subsampled contourlet transform,NSCT)的航拍绝缘子图像边缘提取方法。先利用分段线性灰度变换实现预处理,然后进行NSCT分解,基于分块思想对系数进行分块并求局部阈值,得到边缘图像,最后对边缘检测结果进行形态学滤波使边缘图像更清晰。分别对Lena图像和现场绝缘子图像用Canny算子法、小波模极大值法和所提方法进行图像边缘提取,并对各方法进行性能指标的评价。实验结果验证了所提方法对绝缘子图像边缘检测的有效性,并表明了该方法优于基于Canny算子和小波模极大值的边缘提取方法。Insulator image edge extraction is the important premise of detecting and recognizing insulator defects from aerial images. According to the characteristics of aerial insulator image, a method of aerial insulator image edge extraction based on non subsampled Contourlet transform （NSCT） is proposed in this paper. Firstly, image pretreatment is realized through piecewise linear gray level transformation; then the pretreated image is decomposed using NSCT. The NSCT coefficients are split into blocks, the local threshold value of every coefficient block is calculated, and the binary edge image is obtained based on the threshold. Finally, morphological filtering is used to makes the edge image clearer. Canny operator method, wavelet modulus maximum method and the proposed method were used on Lena image and real field insulator images to carry out image edge extraction; and the performances of different methods are evaluated, and the performance indices are computed. Experimental results demonstrate that the proposed method can obtain better edge extraction effect; and the proposed method is superior and more effective than Canny operator and wavelet modulus maximum methods.

关 键 词：非下采样轮廓波变换 边缘提取 绝缘子图像 系数分块 

分 类 号：TN911.73[电子电信—通信与信息系统] TM726[电子电信—信息与通信工程]