Abstract: ObjectiveLong-distance oil and gas pipelines are critical components of the global energy network. However, complex internal and external environmental factors can easily lead to the formation of defects on pipeline surfaces, posing a serious threat to the structural integrity of the pipelines. To achieve pipeline defect detection, Magnetic Flux Leakage (MFL) inspection technology is widely used for the identification and quantification of pipeline defects. Accurately identifying defect locations using MFL data enables the capture of signal anomalies caused by defects without interrupting normal pipeline operation, facilitating a comprehensive assessment of pipeline integrity. Method To achieve accurate identification of defect locations, an intelligent detection and identification method for pipeline defects is proposed. Baseline calibration and Gaussian denoising are employed for data preprocessing of ultra-high-definition (UHD) MFL in-line inspection signals. An image mapping method is applied to generate pseudo-color maps from the UHD MFL pipeline data, which are then used to construct a pipeline defect identification database. The Detection Transformer (DETR) network is applied to achieve intelligent identification of pipeline defects. Results To validate the effectiveness of the proposed method, pull-through experiments were conducted using multi-diameter UHD MFL in-line inspection tools at a test site in Daxing, Beijing. A total of 341 effective defect identification samples were established. Validation on the test set showed that the DETR model achieved an AP50 of 92.5%, a precision of 90.3%, a recall of 86.3%, and an F1-score of 88.3%. These results indicate that the model can effectively identify pipeline defects and demonstrates significantly superior performance in detecting small-sized and subtle defects compared to the traditional YOLO model. The method was further applied to UHD MFL in-line inspection data from an in-service crude oil pipeline. The results demonstrate the model's capability to identify defects of various shapes and locations within the operational pipeline, with a validated AP50 reaching 84.9%. Conclusion The results indicate that the proposed method can effectively identify defects in in-service oil and gas pipelines, enhances the interpretation efficiency of UHD MFL in-line inspection data, and provides reliable technical support for the remaining strength assessment and remaining life prediction of pipelines containing defects.