Abstract: Due to the influence of interference factors such as stains, light, and shooting angles at the substation site, the instrument information in the images captured by the remote inspection system is weakened, and the pointer information is missing, resulting in a relatively low recognition accuracy of the instruments. Aiming at this problem, the target detection framework of YOLOv5 was adopted, and a cross-fusion feature pyramid network was designed, which enhanced the ability of the target detection network based on YOLOv5 to extract the position information of instruments. To address the problem of pointer fragmentation caused by blurred and tilted instrument images, the U-Net semantic segmentation network is adopted to identify the pointers in the instrument images, achieving the generation of instrument pointers in the interference environment. Experiments show that the proposed intelligent recognition algorithm for sulfur hexafluoride (SF6) instrument based on deep learning has demonstrated strong recognition capabilities in the complex environment of substations, with the accuracy rate of pointer recognition increasing from the original 63% to 96%.

    Key words: substation; remote inspection; deep learning; SF6 instrument; intelligent recognition; cross-fusion; feature pyramid network