Long-distance pipelines are essential for transporting oil and gas, with the quality of welds directly affecting their safety and reliability. Weld defects can emerge during the welding process due to improper techniques or environmental factors, which can result in pipeline leakage or rupture that pose a public safety and environmental risk and can lead to significant economic losses. Therefore, effective weld defect detection is crucial to ensure the safe operation of long-distance pipelines. Although traditional X-ray inspection is commonly used to detect weld defects, it is inefficient and subjective due to its dependence on manual analysis. New developments in computer vision have significantly improved the efficiency and accuracy of automated technology for defect recognition in pipeline weld X-ray images. As a result, there is an urgent need for a comprehensive review to support the development of this field and provide valuable insights. This paper comprehensively evaluates the progress in the technology for the intelligent recognition of defects in pipeline weld X-ray images, focusing on preprocessing and defect detection techniques. This review explores three key directions for intelligent weld defect recognition: signal processing, feature design, and deep learning-based methods. Deep learning-based defect recognition techniques were examined in detail from five primary perspectives: dataset creation, image classification, semantic segmentation, object detection, and performance evaluation. Finally, the challenges and future development trends in the intelligent recognition of defects in pipeline weld X-ray images are discussed, emphasizing areas that require further research and innovative advancements.