The invention discloses a substation disconnecting switch detection and identification method based on Faster RCNN. The method comprises the following steps of obtaining a monitoring image of a substation; inputting the monitoring image into a Faster RCNN judgment model, and obtaining a possible target window of the disconnecting switch; and obtaining a disconnecting switch window from the possible target window through a classification identification network. According to the method, feature extraction is carried out on the transformer substation monitoring image through the deep neural network, manual participation is reduced, the problem of misjudgment caused by a traditional image processing method is avoided, and errors and non-comprehensiveness of traditional manual feature extraction are avoided. The characteristics of the isolation switch extracted by using the CNN network have relatively high robustness, relatively good retrieval precision can still be kept when the backgroundof the retrieval image is complex or the monitoring angle is changed, and the positioning and identification of the disconnecting switch are carried out in the same model, so that the retrieval efficiency of key equipment is greatly improved.