Infrared and visible image matching method for transmission line power equipment based on local and global features of PIIFD+MA-SCW

Abstract To improve the accuracy of infrared and visible image matching of transmission line power equipment, this paper proposed an infrared and visible image matching method of transmission line power equipment based on local and global features of PIIFD + MA-SCW to improve the accuracy of image matching. First, the infrared and visible images were grayscale processed, denoised, and enhanced, and the feature points were detected by multi-scale corner detection algorithm to realize feature extraction. Then, the Partial Intensity Invariant Feature Descriptor (PIIFD) and multi-angle partitioned for the Shape Context Weighted descriptor (MA-SCW) of each feature point were constructed. Subsequently, a similar bidirectional matching method was used to complete preliminary matching. Finally, the Random Sample Consensus (RANSAC) method was used to obtain the correct matching point pairs to achieve the final matching. The experimental results show that the image matching method based on local and global features of PIIFD + MA-SCW exhibits excellent matching performance. Simultaneously, this method is robust when the image undergoes scale changes, rotation changes, and local feature differences. This research is of great significance for realizing better analysis and monitoring of transmission line power equipment and for maintaining the stability of power systems.