A prohibited object detection and recognition system for X-ray images
Yang Zheng, Jie Liu, Shuwu Zhang · 2021
The object detection and recognition system for X-ray images containing prohibited object divided into two parts: The construction of synthetic data containing prohibited objects, as well as the detection and recognition of prohibited objects. Among them, the construction of prohibited object data includes object classification, attribute analysis, object image acquisition, fusion of object image and X-ray image. After the above steps, the X-ray images with prohibited objects such as knives and guns are similar to the real images are obtained, which effectively solves the problem of insufficient X-ray image. In the process of detection and recognition of prohibited objects, by changing the color and shape of training data, the data source is expanded, which makes the system have higher robustness and improves the detection efficiency of prohibited objects. The detection and recognition system can effectively complete the real-time detection and recognition of knives and guns in luggage, save a lot of manpower and material resources, and ensure the safety of road, railway and air transportation.