X-ray safety check image structural description method and system

Abstract

The invention discloses an X-ray safety check image structural description method and system; the method comprises the following steps: training X-ray images of existing various prohibited articles so as to form a corresponding prohibited article X-ray image training image database; further segmenting a to-be-identified X-ray image on previous basis; comparing and coupling the segmented image with the trained images in the training image database; analyzing and parsing the X-ray safety check image according to the comparison and coupling result, thus forming structured description data. The method and system are wide in application scopes, stable and reliable in performance, can reduce artificial monitoring cost, thus improving prohibited articles discovery probability and accuracy, furthering improving X-ray safety check intelligentize and informatization level, and effectively solving existing technical problems.

Description

technical field [0001] The invention relates to security technology, in particular to X-ray security inspection technology. Background technique [0002] In recent years, with the rapid development of aviation, railway, urban transportation and express logistics industry, X-ray security inspection machines are widely used to ensure the safety of passengers and goods. The purpose of X-ray security inspection is to use X-ray perspective images to find or discover prohibited items contained in cargo packages, mainly including flammable, explosive, corrosive items, controlled knives and firearms, rare animal and plant products, drugs and precursors. Toxic chemicals, regional exclusive commodities, and items that need to be investigated by the public security. [0003] At present, X-ray security inspection machines mainly rely on manual monitoring to work. The monitoring personnel observe X-ray images and identify suspicious items in the images according to the shape and color o...

AI Technical Summary

Problems solved by technology

The identification technology adopted by these X-ray security inspection machine products can identify some simple prohibited items such as lithium batteries, metal guns, metal knives, lighter gas tanks, etc., but the identification technology adopted by these X-ray security inspection machines generally has the following problems: The types of prohibited items are limited, and the accuracy rate is not high. It is especially limited to the X-ray machines of the company. It cannot support other similar X-ray machines. It does not have a learning function. Users cannot add samples of prohibited items by themselves, nor can they give the alarm items details

Although the system realizes basic image segmentation and detection and recognition functions, the accuracy of segmentation and recognition is not high, and it is also unable to give a description of the nature of contraband, which is far from practical application.

[0009] To sum up, the existing X-ray security inspection technology cannot solve the problem of identification and description of X-ray images

Furthermore, the current X-ray machine identification and alarm technology is only for a small number of items such as bombs, guns, knives, batteries, etc., and the database of prohibited items cannot be added. Query and retrieval function, unable to analyze, manage and apply information to massive X-ray images

Images