Abstract: When recognized at the same time, the digital and pointer dual dial images of lightning arrester monitor captured by existing machine vision have the issue of low accuracy. This paper presents an intelligent identification method IIAM, which can accurately identify the dual dial of lightning arrester monitor. The target detection was carried out for the arrester monitor through model YOLOv4, and the instrument area was divided into digital and pointer dial area. The pointer dial area was processed through semantic segmentation, linear array compression and so on. Aiming at the abnormal situation of card digital display in the digital dial area, this paper proposed a text detection and recognition technology based on OCR, so that accurate readings were obtained through extraction and procession of abnormal data. The test shows that the recognition accuracy of digital dial area is 96.5%, and the pointer dial area is 95.5% under normal exposure. In the case of overexposure and lack of exposure, the recognition accuracy of pointer and digital dial reaches more than 90%.