It is one of the important tasks to maintain the safe and stable operation of the power grid to discover equipment defects or hidden safety hazards in time through daily inspections and take necessary measures to deal with them. Currently, power grid companies are exploring intelligent substation inspection technology based on image recognition to reduce the work pressure of operation and maintenance personnel and improve the quality and efficiency of operation and maintenance. Metal corrosion is one of the most frequent appearance defects of substation equipment. In severe cases, it will cause problems such as overheating of the equipment and breakage of the joints. This paper proposes an improved object detection network based on image recognition technology. On the basis of the SSD algorithm, the ResNet network is used instead of the VGG network, and the attention mechanism is added. Experiments show that the algorithm can effectively identify the rusty area of the equipment. Applying this algorithm to the intelligent inspection work of substations will effectively improve the timeliness and accuracy of defect discovery.