论文部分内容阅读
针对电力设备目标定位问题,研究了基于改进的尺度不变特征转换(scale invariant feature transform,SIFT)精确图像配准的定位方法。首先用改进的SIFT特征描述子提取图像中的特征点,降低了特征向量描述子的维数,大大提高了算法的速度;然后采用欧氏距离对特征点进行初始匹配,由于初始匹配过程中存在误匹配,采用改进的随机取样一致性(random sample consensus,RANSAC)算法对阈值进行自动调整,消除了错误匹配;最后,以电力系统中的刀闸、变压器为例,采用旋转、缩放、光照变化及加噪声图像验证了该算法。实验结果表明:改进后的算法不仅继承了SIFT算法的鲁棒性,而且提高了算法的速度和匹配精度,可以较好地应用于电力设备目标定位中。
In order to solve the problem of target location of power equipment, a method of locating precise image registration based on improved scale invariant feature transform (SIFT) is studied. Firstly, the improved SIFT feature descriptor is used to extract the feature points in the image, which reduces the dimension of the feature vector descriptor and greatly increases the speed of the algorithm. Then, the Euclidean distance is used to initial match the feature points. Since the initial matching process exists The error threshold is adjusted automatically by the improved random sample consensus (RANSAC) algorithm to eliminate the mismatch. Finally, taking the switch and transformer in the power system as an example, the rotation, zooming and lighting changes The noise-enhanced image validates this algorithm. The experimental results show that the improved algorithm not only inherits the robustness of the SIFT algorithm, but also improves the speed and matching accuracy of the algorithm and can be well applied to the target location of power equipment.