Power transmission lines are vital infrastructures in our daily lives. We have developed a remote-controlled robot that can run on the transmission lines and capture clear images to improve the safety, accuracy, and efficiency of power transmission line inspections. On the other hand, anomaly detections in the images of the power transmission lines are also a difficult task for humans because they have to inspect all the captured images thoroughly. We propose a method of photographing corrosion products generated on the surface of ground wires by using photoluminescence. In addition, we propose an anomaly detection method using deep neural networks further to improve the accuracy and efficiency of the inspections. In this paper, we present the results of field tests conducted to verify the essential operation of the inspection robot and the anomaly detection methods.