基于红外成像技术的电气设备过热区域自动定位.doc

基于ASIFT算法的绝缘子视频图像的识别与定位* 崔克彬1，苑津莎2，李宝树3 （1.华北电力大学 计算机系，保定 071003；2. 华北电力大学 电子系，保定 071003；3.华北电力大学 电力系，保定 071003） 摘要：为了实现智能提取直升机巡检视频中的绝缘子图像，基于ASIFT原理的图像处理技术和数据库技术，提出了一种新的绝缘子图像识别与定位方法。该方法首先建立标准的绝缘子图库，通过改进UL- PCNN红外图像分割算法提取绝缘子特征值，然后将输电线路视频与标准图库中的绝缘子图片利用ASIFT算法进行匹配，进而识别和定位视频中的绝缘子。实验结果表明，ASIFT 方法具备良好的抗绝缘子图像仿射变形性能，可以在少量人工辅助的条件下对图像进行处理，提高了架空输电线路绝缘子故障检测的自动化处理程度。 关键词：绝缘子；航拍视频图像；ASIFT算法；识别与定位 Identification and Location of Insulator Video Images Based on ASIFT Algorithm Cui Kebin1 , Yuan Jinsha2, Li Baoshu3 (1. Department of Computer , North China Electric Power University, Baoding, 071003. 2.Department of electronics , North China Electric Power University, Baoding, 071003 3.Department of Electrical Engineering , North China Electric Power University, Baoding, 071003) Abstract: A new insulator image identification and location method was proposed. Based on Image processing technology and database technology of the principle of ASIFT, it can be intelligent extraction helicopter patrol in the video image of insulators. Firstly, the establishment of a standard insulator gallery, extracting characteristic values by improving the insulator UL-PCNN infrared image segmentation algorithm, and then let transmission line captured video of patrol matching with a standard library of insulator image by ASIFT, and then identify and positioning the video insulators, ASIFT image recognition method simulate the longitude and latitude angle camera angle optical axis and match image features in the affine space, the results show, ASIFT method has good anti-affine deformation properties. This method can be processed the image at a small image artificial conditions, and improved automation of image processing based on overhead transmission line insulator fault detection. Keywords: insulator; Aerial video; ASIFT Algorithm; Identification and Location 1 引言 绝缘子是架空输电线路的重要组成部分，由于长期暴露在野外，受到持续的机械张力、电气闪络、材料老化影响导致绝缘水平下降如不及时修复更换，，造成大面积的停电和巨大的经济损失[1-2]，因此对其所拍摄的视频进 行处理有着很重要的实际意义。 目前国内外学者通过图像处理技术对绝缘子的状态监测与识别