Autonomous and intelligent security checking has always been the research hotspot in industry and academia. Intelligent analysis of X-ray security checking images based on computer vision is an important means to promote the implementation of intelligent security checking technology due to its portability, easy maintenance, and low cost. We noticed that DETR-based object detection algorithms exhibit strong performance on various datasets. Compared with the object detection algorithms of the CNN structure, DETR uses the Transformer-based feature extraction structure to achieve effective feature aggregation through global attention. In this paper, we explore in depth the application of the DETR algorithm to the detection of prohibited items in X-ray security checking images. Some CNN based object detection models, such as YOLO, SSD et al., have taken into the comparison. The simulation results based on SIXray dataset have demonstrated the state-of-the-art performance of DETR model in prohibited item detection in X-ray images.