CN-122016888-A - Comprehensive liquid qualitative method, X-ray security inspection equipment and device

CN122016888ACN 122016888 ACN122016888 ACN 122016888ACN-122016888-A

Abstract

The invention relates to the technical field of security inspection and discloses a comprehensive liquid qualitative method, X-ray security inspection equipment and a device, wherein the method comprises the steps of obtaining a transmission image and a back scattering image of an object to be inspected, and establishing a coordinate association relation of the same association container in the transmission image and the back scattering image; the method comprises the steps of respectively carrying out liquid segmentation on a transmission image and a back scattering image to correspondingly obtain a non-shielding liquid profile and a liquid profile, extracting liquid region original data and container shell region original data, calculating to obtain transmission liquid characteristics, back scattering liquid characteristics, transmission container characteristics and back scattering container characteristics based on the liquid region original data and the container shell region original data, fusing and constructing a liquid characteristic matrix, and obtaining a liquid attribute classification result through a liquid attribute discrimination model.

Inventors

LI LEI
SONG CHUNTIAN
ZHU YANGMIN
YANG HAIYING
FAN YU
JIA XIAOWEN

Assignees

山东鲲勋安检科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (10)

1. An integrated liquid qualitative method, the method comprising: Acquiring a transmission image and a back scattering image of an object to be detected, and establishing a coordinate association relation of the same association container in the transmission image and the back scattering image based on the transmission image and the back scattering image; Respectively carrying out liquid segmentation on the transmission image and the back scattering image to correspondingly obtain a non-shielding liquid contour and a liquid contour, and extracting liquid region original data and container shell region original data based on the non-shielding liquid contour and the liquid contour; based on the liquid region original data and the container shell region original data, calculating to obtain transmission liquid characteristics, back scattering liquid characteristics, transmission container characteristics and back scattering container characteristics, and fusing to construct a liquid characteristic matrix; and inputting the liquid characteristic matrix into a preset liquid attribute distinguishing model to obtain a liquid attribute classification result.
2. The method of claim 1, wherein establishing the coordinate association of the same association container in the transmission image and the back-scattered image based on the transmission image and the back-scattered image comprises: Spatially registering the transmission image and the back scattering image to obtain registered coordinates; Respectively carrying out container detection and segmentation on the transmission image and the back scattering image, and identifying a container region to obtain a projected image container coordinate and a back scattering container result frame, wherein the container region comprises a container contour and a surrounding frame; And mapping the container coordinates in the transmission image to a back scattering coordinate system by using the registered coordinates, generating a transmission container mapping frame, and establishing a coordinate association relation of the same association container in the transmission image and the back scattering image based on the transmission container mapping frame and a back scattering container result frame.
3. The method of claim 2, wherein spatially registering the transmission image with the back-scattered image to obtain registered coordinates comprises: Mapping coordinates in the transmission image to coordinates in the back scattering image by adopting a preset translation coefficient and a preset scaling coefficient, and expressing a coordinate formula after registration as follows: ; ; Wherein, the 、 Is the translation coefficient of the transmission image to the back-scattered image map, 、 Is the scaling factor of the transmission image to backscatter image mapping, Representing the transmission point coordinates in the transmission image.
4. The method of claim 2, wherein establishing a coordinate association of the same association container in the transmission image and the backscatter image based on the transmission container mapping frame and the backscatter container result frame comprises: and calculating the cross-over ratio and the complete cross-over ratio between the transmission container mapping frame and the back scattering container result frame, judging the same container if the cross-over ratio and the complete cross-over ratio are both larger than a preset threshold value, and establishing a pair of association relations between the same container and the back scattering container coordinates in a transmission image coordinate system.
5. The method of claim 1, wherein the step of determining the position of the substrate comprises, Respectively carrying out liquid segmentation on the transmission image and the back scattering image to correspondingly obtain a non-shielding liquid contour and a liquid contour, and extracting liquid area original data and container shell area original data based on the non-shielding liquid contour and the liquid contour, wherein the liquid segmentation method comprises the following steps: liquid segmentation is carried out on the container in the transmission image, the non-occlusion liquid outline is extracted, and the non-occlusion liquid outline is matched with the associated container; liquid segmentation is carried out on the container in the back scattering image, the liquid outline is extracted, and the liquid outline is matched with the associated container; Corresponding liquid region raw data are respectively extracted from transmission raw data and back-scattering raw data which are acquired in advance, and meanwhile container shell region raw data are extracted from the back-scattering raw data.
6. The method of claim 1, wherein the transmissive liquid characteristic comprises a first mean, a first standard deviation, a first quartile Q1, a first quartile Q3, a first skewness, and a first coefficient of variation; The back scattering liquid characteristic comprises a second mean, a second standard deviation, a second quartile Q1', a second quartile Q3', a second skewness and a second coefficient of variation, and the back scattering container characteristic comprises a container mean and a back scattering liquid to container ratio.
7. The method according to claim 1, wherein the method further comprises: And constructing a shielding characteristic matrix based on the transmission liquid characteristic, the back scattering liquid characteristic, the transmission container characteristic and the back scattering container characteristic, and inputting the shielding characteristic matrix into a preset shielding judgment model to obtain a liquid shielding state judgment result.
8. The method of claim 7, wherein the transmission image comprises a transmission high-energy image and a transmission low-energy image, the method further comprising: calculating a pseudo color image based on the transmission high-energy image and the transmission low-energy image, wherein the pseudo color image comprises R, G, B channels; Fusing three channels of the transmission high-energy image, the transmission low-energy image and the transmission pseudo-color image with the back scattering image after registration mapping to generate multi-channel fused image data; inputting the multichannel fusion data into a preset classification model, and synchronously obtaining classification results of liquid properties and shielding states, wherein the classification results of the liquid properties and the shielding states are used for supplementing or verifying the classification results of the liquid properties and the judgment results of the liquid shielding states.
9. An X-ray security inspection device is characterized by comprising an X-ray transmission imaging module, a Compton back scattering imaging module, an image feature extraction module and a first liquid attribute distinguishing module; the X-ray transmission imaging module is used for acquiring a transmission image of an object to be detected; the Compton back scattering imaging module is used for back scattering images of the to-be-detected object; The image feature extraction module is used for executing the steps of acquiring a transmission image and a back scattering image of an object to be detected in the claim 1, establishing a coordinate association relation of the same association container in the transmission image and the back scattering image based on the transmission image and the back scattering image, respectively carrying out liquid segmentation on the transmission image and the back scattering image to correspondingly obtain a non-shielding liquid contour and a liquid contour, extracting liquid area original data and container shell area original data based on the non-shielding liquid contour and the liquid contour, calculating to obtain transmission liquid features, back scattering liquid features, transmission container features and back scattering container features based on the liquid area original data and the container shell area original data, and fusing and constructing a liquid feature matrix; The first liquid attribute distinguishing module is configured to execute the step of inputting the liquid feature matrix into a preset liquid attribute distinguishing model to obtain a liquid attribute classification result in claim 1.
10. An integrated liquid characterization device, the device comprising: The transmission image and back scattering image association relation establishing module is used for acquiring the transmission image and the back scattering image of the object to be detected and establishing the coordinate association relation of the same association container in the transmission image and the back scattering image based on the transmission image and the back scattering image; The liquid contour and container shell extraction module is used for respectively carrying out liquid segmentation on the transmission image and the back scattering image, correspondingly obtaining a non-shielding liquid contour and a liquid contour, and extracting liquid region original data and container shell region original data based on the non-shielding liquid contour and the liquid contour; The liquid characteristic matrix construction module is used for calculating transmission liquid characteristics, back scattering liquid characteristics, transmission container characteristics and back scattering container characteristics based on the liquid region original data and the container shell region original data, and fusing and constructing a liquid characteristic matrix; The second liquid attribute distinguishing module is used for inputting the liquid characteristic matrix into a preset liquid attribute distinguishing model to obtain a liquid attribute classification result.

Description

Comprehensive liquid qualitative method, X-ray security inspection equipment and device Technical Field The invention relates to the technical field of security inspection, in particular to a comprehensive liquid qualitative method, X-ray security inspection equipment and a device. Background Currently, the case detection is an X-ray transmission technology, and the effective atomic number of a detected object is calculated and the substances are divided into three main categories by pseudo-color through scanning and acquiring high-energy and low-energy images of the detected object, namely orange organic substances, green mixtures and light metals, and blue inorganic substances. And because the objects in the security inspection are in a substance superposition form, the boundary between the mixture and the light metal and the organic or inorganic substances is relatively fuzzy, when the organic substances and the non-organic substances are superposed, the color is in a green and orange or blue interweaving state, the composition of the mixture is difficult to accurately analyze, when the thickness of the organic substances is too high, the color is also in the interweaving state, and the detection precision of the non-overlapped organic substances is limited. For a conventional security inspection machine, the X-ray transmission is difficult to finish the organic matter refined analysis, and for a security inspection CT device, although the material density distribution calculation can be realized, the material attribute still cannot be accurately determined, and the speed is low, so that the requirement of rapid security inspection is difficult to meet. The security inspection is to judge the category of the detected object according to the shape and color of the detected object, and the material attribute cannot be determined. Therefore, the existing transmission security inspection technology can only judge the shape of the detected object, but can not effectively identify the organic matters such as liquid, explosive, drugs and the like, so that the problems of incapability of qualitatively identifying the object, insufficient identification precision and the like are caused, and the problem of difficulty in judging the material property is caused. Disclosure of Invention The invention provides a comprehensive liquid qualitative method, X-ray security inspection equipment and a device, which are used for solving the problems that qualitative identification cannot be carried out on articles, identification precision is insufficient and the like, and the quality of the materials is difficult to judge. In a first aspect, the present invention provides an integrated liquid characterization method, comprising: Acquiring a transmission image and a back scattering image of an object to be detected, and establishing a coordinate association relation of the same association container in the transmission image and the back scattering image based on the transmission image and the back scattering image; Respectively carrying out liquid segmentation on the transmission image and the back scattering image to correspondingly obtain a non-shielding liquid contour and a liquid contour, and extracting liquid region original data and container shell region original data based on the non-shielding liquid contour and the liquid contour; Based on the original data of the liquid area and the original data of the container shell area, calculating to obtain a transmission liquid characteristic, a back scattering liquid characteristic, a transmission container characteristic and a back scattering container characteristic, and fusing to construct a liquid characteristic matrix; inputting the liquid characteristic matrix into a preset liquid attribute distinguishing model to obtain a liquid attribute classification result. According to the comprehensive liquid qualitative method provided by the invention, through fusing the X-ray transmission and back scattering dual-mode imaging data, a complete processing flow from image registration, container association and liquid segmentation to multi-feature extraction and modeling is constructed, so that the high-precision discrimination of liquid properties in the container is realized, namely, the liquid features in the transmission and back scattering data and the container features overlapped with the liquid are comprehensively extracted, the attribute discrimination of target organic liquid is completed, the accuracy of substance qualitative is improved, the limitation that the traditional single transmission technology has low resolution ratio on organic matters and can not distinguish overlapped substances is effectively overcome, and the automatic identification capability and reliability of liquid dangerous goods in a security inspection scene are improved. In an alternative embodiment, establishing the coordinate association relationship of the same association container in the