Automatic Detection and Identification of Controlled Knives Based on Improved SSD Model

The automatic detection and identification of security inspection system is the inevitable trend of intelligent development. It is an important research problem to improve the accuracy and efficiency of X-ray security inspection equipment to detect dangerous goods. Taking the automatic tool detection and recognition system of security inspection system as the research object, aiming at the fact that the shallow characteristic chart representation ability of SSD algorithm is not strong, the small target feature of training stage gradually disappears, and the detection accuracy is not high, and there are some problems, such as missing detection and false detection of small targets such as controlling tool in security inspection dangerous goods. This paper proposes to replace the basic network VGGI 6of SSD with ResNetl01 with stronger anti-degradation performance to construct SSD-ResNet101 network structure. On this basis, the feature fusion method is used to make the shallow layer merge the deeper features to increase the receptive field of the shallow feature map. Make full use of the context information to improve the detection accuracy of the small target of dangerous goods in security inspection. The analysis of experimental test data shows that the improved algorithm has high detection accuracy, good real-time performance and good robustness to small targets such as control tools in security picture.