The invention relates to the technical field of computer image processing and recognition, and more specifically relates to a cable equipment temperature abnormity location and recognition method. Firstly, the training images of the target quantity are generated by expanding the samples together with the original images to act as training samples on the basis of the images taken by a cable tunnelinspection system, and the Faster R-CNN network model is constructed to train the training samples to obtain the target detection network, the position of the connector area in the visible light photograph is obtained and mapped to the infrared photograph, and the temperature of the cable connector is analyzed so as to alarm the abnormal situation in time. The invention selects to train the FasterR-CNN target detection network parameters, the RPN network is applied to extract the pre-selected area to train the target detection network, and the capacity of the convolutional neural network forextracting the two-dimensional image features is fully utilized to realize cable equipment temperature abnormity location and recognition so as to have high accuracy, wide applicability, good recognition quality and high recognition speed.