| Now X-ray screening devices can only classify substances for detection,such as organic substances,inorganic substances,metals,and liquids.Functional classification of detection substances still need security personnel to carry out visual inspection and identification.Due to huge traffic,visual discrimination ability and some different factors,often occur undetected phenomenon of dangerous goods,especially metal-like substances,Thus in this study will automatically identify and identify the cutleries in the security inspection,reduce the occurrence of missed detection errors,and reduce the work pressure of security staff.This subject first uses R-value algorithm to extract the metal material image from the X-ray safety inspection image,generates preprocessed image,and performs edge detection on the image to extract its finished edge.According to the characteristics of X-ray image detection,the wavelet decomposition and reconstruction method is used to extract the edges of the image,and the complete edges can be extracted.Then the morphology of the image edges is processed,and the connected regions are marked to realize image segmentation.Through the comparison of the imaging effects of the cutlery and the non-cutlery object under X-ray,the shape cutlery feature element is defined,and the feature element extraction is completed according to the cutlery morphological feature extraction method of this subject so as to carry out subsequent pattern recognition judgment.Then,the method of identifying and classifying the cutlery-defined feature elements is selected for this task.According to the characteristics of the defined features,the classification methods such as BP neural network,LVQ neural network,decision tree,and random forest,as well as the combination of decision tree and random forest proposed in this paper are used to compare the results of feature element classification.The comparison results show that the combination of decision tree and random forest can reduce the tool recognition error rate.In the process of building the automatic cutlery identification system,it is based on the VS2013 compilation environment and the MATLAB simulation program to implement the system functions.And ZYNQ as the main chip equipped with FreeRTOS real-time operating system,designed X-ray security equipment,data acquisition and transmission system to improve the formation of grayscale image resolution,expand the image grayscale range,in order to improve the accuracy of image data.Finally,in order to visually display the detection effect of the cutlery recognition verification system,an MFC function library is used to build the display interface,and the original image and the recognition result of the actual detection object are displayed,and a cutlery is identified to give an alarm prompt.The use of cutlery identification systems can increase the efficiency of safety inspections in public places,reduce the chance of subjective identification errors by security personnel,and achieve objectivity and automation requirements for safety inspections. |