Abstract
In order to quickly and accurately identify and locate dangerous goods in security inspection images, this article proposes a real-time method to detect and locate dangerous goods by improving Faster RCNN algorithm and applies it to security inspection machine system. The main design idea of the system is to prepare the security inspection image data set and build an improved Faster RCNN model using Caffe framework. Through training and testing of the model, the recognition model with an average recognition rate of more than 80%, especially the detection effect of small targets, has been greatly improved. Finally, the model is applied to the current security inspection system, and an intelligent security image recognition system is designed and implemented.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have