基于SIFT特征匹配的电力设备图像变化参数识别

Image Variation Parameter Recognition of Electrical Power Equipment Based on SIFT Feature Matching

在变电站视频监视系统中,可通过视频巡检来及时发现某些异常状态,以免产生故障。图像的某些参数的改变是判别图像状态改变的重要依据,提出了一种基于SIFT,OTSU和RANSAC相结合的特征匹配的电力设备图像变化参数识别算法。先对样本资料图像和监测图像进行SIFT特征匹配,与OTSU相结合消除干扰匹配特征点,再通过RANSAC随机抽样一致性算法消除错误匹配特征点,根据匹配结果识别电力设备图像变化角度和缩放比例这2个重要参数。仿真实验表明,此算法简单易行,精度高,可用于电力铁塔倾斜角度和指针式仪表指针旋转角度的识别,也可用于视频巡检中缩放系数的识别。

In video monitoring system of substation,in-process video inspection is used to detect abnormalities in timely so as to avoid failures.Changes in certain parameters of the image constitute an important basis in identifying the changes in image status.A power installation image change parameter identification algorithm based on SIFT,OTSU and RANSAC feature matching was presented.Firstly,a SIFT feature matching was performed on the sample image and monitoring image.Then the interference matching feature points were eliminated in combination with OTSU,and the RANSAC random sampling agreement algorithm was used to eliminate the wrong matching feature point.Finally,two important parameters including the variation angle and zoom coefficient of power installation image were identified based on the matching result.As is proved in the simulation test,this algorithm features simplicity and high precision and it can be used in recognizing the electric power tower inclination angle and meterneedle rotation angle.Meanwhile it can be used in recognizing the zoom coefficient in video inspection as well.