WO2017101514A1 - 检查货物的方法、系统和装置 - Google Patents

检查货物的方法、系统和装置 Download PDF

Info

Publication number
WO2017101514A1
WO2017101514A1 PCT/CN2016/097579 CN2016097579W WO2017101514A1 WO 2017101514 A1 WO2017101514 A1 WO 2017101514A1 CN 2016097579 W CN2016097579 W CN 2016097579W WO 2017101514 A1 WO2017101514 A1 WO 2017101514A1
Authority
WO
WIPO (PCT)
Prior art keywords
suspicious
area
atomic number
region
suspect
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2016/097579
Other languages
English (en)
French (fr)
Inventor
李元景
赵自然
刘耀红
王启立
李强
顾建平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Nuctech Co Ltd
Original Assignee
Tsinghua University
Nuctech Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University, Nuctech Co Ltd filed Critical Tsinghua University
Priority to SG11201709757VA priority Critical patent/SG11201709757VA/en
Priority to BR112017025624-0A priority patent/BR112017025624B1/pt
Publication of WO2017101514A1 publication Critical patent/WO2017101514A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/06—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption
    • G01N23/10—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption the material being confined in a container, e.g. in a luggage X-ray scanners
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/04—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/40—Imaging
    • G01N2223/401—Imaging image processing
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/60—Specific applications or type of materials
    • G01N2223/626—Specific applications or type of materials radioactive material
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10116—X-ray image

Definitions

  • the present disclosure relates to automatic detection of suspected objects in a radiation image, and more particularly to methods and systems for inspection of high atomic number materials in large container scanning systems.
  • the present disclosure proposes a method and system for inspecting goods.
  • a method of inspecting a cargo comprising the steps of: obtaining a transmission image of the inspected cargo; processing the transmission image to obtain a suspicious region; extracting a local texture feature of the suspicious region, and utilizing The previously created model classifies the local texture features of the suspicious region to obtain a classification result; extracts the contour shape feature of the suspicious region, and compares the contour shape feature with a previously created standard template to obtain a comparison a result; and synthesizing the classification result and the comparison result, determining that the suspicious region contains a high atomic number substance.
  • the method further includes the steps of: acquiring equivalent atomic number information of the suspect region; wherein the combining step integrates the classification result, the comparison result, and equivalent atomic number information to determine The suspicious region contains high atomic number materials.
  • the step of processing the transmission image to obtain a suspicious region comprises: performing contrast enhancement on the transmission image; performing multiple binarization operations on the contrast-enhanced transmission image; and determining the binarization The connected regions of the transmitted image and determine the boundaries of the connected regions.
  • the step of extracting the contour shape feature of the suspect region includes: determining an outline of the suspect region using an edge extraction algorithm to obtain a ( ⁇ , ⁇ ) characteristic curve of the contour.
  • the step of comparing the contour shape feature to a previously created standard template comprises calculating a distance and magnitude consistency between a ( ⁇ , ⁇ ) characteristic curve of the contour and a standard template And a ⁇ hop value of the ( ⁇ , ⁇ ) characteristic curve of the contour; comparing the distance, the amplitude consistency, and the ⁇ hop value to a threshold value, respectively.
  • the edge feature extraction algorithm is used to obtain the contour line of the suspect object in the suspicious area, and then the centroid of the area enclosed by the contour line is calculated, and then the calculated centroid is used as the coordinate origin to obtain the contour line and the coordinate.
  • the method further comprises the step of highlighting a region of the fluoroscopic image that is considered to be a high atomic number material.
  • a system for inspecting goods comprising: an X-ray scanning device that obtains a transmission image of the inspected goods; and a data processing device that processes the transmission image to obtain a suspicious area, extracting the a local texture feature of the suspect area, and classifying the local texture features of the suspect area by using a model created in advance, obtaining a classification result, extracting a contour shape feature of the suspect area, and extracting the contour shape feature and the
  • the created standard templates are compared, the comparison results are obtained, and the classification results and the comparison results are combined to determine that the suspicious region contains a high atomic number material.
  • an apparatus for inspecting a cargo comprising: a device for processing a transmission image of an inspected cargo to obtain a suspicious region; extracting a local texture feature of the suspicious region, and utilizing a model created in advance And classifying the local texture features of the suspicious region to obtain a classification result; extracting the contour shape feature of the suspicious region, and comparing the contour shape feature with a previously created standard template to obtain a comparison result Means; and integrating the classification result and the comparison result to determine a device containing the high atomic number substance in the suspect region.
  • FIG. 1A and 1B illustrate a schematic structural view of a cargo inspection system in accordance with an embodiment of the present disclosure
  • FIG. 2 is a schematic flow chart describing a cargo inspection method according to an embodiment of the present disclosure
  • FIG. 3 is a schematic flow chart describing a process of creating a classification model in a solution according to an embodiment of the present disclosure
  • FIG. 4 is a schematic flow chart describing a method of checking using a created model in a dual energy mode according to an embodiment of the present disclosure
  • FIG. 5 is a schematic flow chart depicting a single energy mode in accordance with an embodiment of the present disclosure
  • FIG. 6 is a schematic diagram depicting a model of a high atomic number material used in a scheme in accordance with an embodiment of the present disclosure
  • 7A, 7B, 7C, 7D, 7E, 7F, 7G, 7H, and 7I are ( ⁇ , ⁇ ) characteristic curves describing the use of a high atomic number material model in a scheme according to an embodiment of the present disclosure. Schematic diagram.
  • the present disclosure is directed to the problem of failing to accurately and accurately determine a high atomic number substance in a cargo in the prior art. It is proposed to determine whether the suspicious region contains a high atomic number substance based on local texture features and/or contour shape features of the suspicious region in the inspected cargo.
  • the shape feature of the suspicious object is described by the ( ⁇ , ⁇ ) feature
  • the texture feature of the suspicious object is described by the LBP (Local Binary Pattern) feature.
  • the effective atomic number information can be further used in other embodiments (eg, dual energy systems) for further determination. This makes it possible to use the texture features of the image, the edge information of the suspicious object and/or the material information to simplify the high atomic number material detection process, shorten the calculation time and reduce false alarms.
  • FIGS. 1A and 1B are schematic structural views of an inspection system according to an embodiment of the present disclosure.
  • FIG. 1A shows a top plan view of the inspection system
  • FIG. 1B shows a front view of the inspection system.
  • the radiation source 110 generates X-rays that are collimated by the collimator 120 to perform a security check on the moving container truck 140
  • the detector 150 receives radiation that penetrates the truck, such as a computer.
  • the data processing device 160 of the class obtains a transmission image, and processes the transmission image to obtain a determination result.
  • FIG. 2 is a schematic flow chart describing a cargo inspection method according to an embodiment of the present disclosure.
  • a transmission image of the container truck 140 is obtained by scanning in step S21, and then the transmission image is processed by the data processing device 160 in step S22 to obtain a suspicious area.
  • step S23 the local texture features of the suspicious region are extracted, and the local texture features of the suspicious region are classified by using a model created in advance to obtain a classification result.
  • step S24 the data processing device 160 extracts the outline shape feature of the suspicious area, and compares the outline shape feature with a previously created standard template to obtain a comparison result.
  • step S25 the classification result and the comparison result are combined to determine that the suspicious region contains a high atomic number substance.
  • FIG. 3 is a schematic flow chart describing a process of creating a classification model in a scheme according to an embodiment of the present disclosure.
  • a high atomic number substance image is input, and then at step S32, local texture features of the image, such as LBP features, are extracted.
  • the LBP feature is an operator for describing the local texture of an image, which has the advantages of rotation invariance and gray invariance.
  • nuclear materials, nuclear weapons, radioactive dispersion devices and their shielding devices have a regular shape such as a cube, a sphere or a cylinder. A large number of suspicious regions can be obtained during the image binarization of the connected regions, and substantially all of the high atomic number simulation materials are contained in a small portion of the region.
  • FIG. 4 is a schematic flow chart describing a method of checking using a created model in a dual energy mode in accordance with an embodiment of the present disclosure.
  • FIG. 5 is a schematic flow chart depicting a single energy mode in accordance with an embodiment of the present disclosure.
  • the input of the high atomic number automatic detection can be either a fluoroscopic image or a fluoroscopic image and a material map (equivalent atomic number map).
  • the high atomic number detection is performed according to the flow shown in FIG.
  • the detection is performed in accordance with the flow shown in FIG.
  • the equivalent atomic number information of the inspected goods can be obtained by high-energy and low-energy images, and the information can be used together with the shape information to determine whether the image contains high atomic number substances.
  • a fluoroscopic image is input.
  • the input image is then preprocessed.
  • the container area in the image is detected such that subsequent processing is only performed on the container image, the amount of data is reduced, and then the contrast enhancement is performed on the image in step S403.
  • the cargo area can be obtained after the container body detection algorithm in the pre-processing stage for subsequent high atomic number substance detection.
  • the image may be subjected to logarithmic transformation or logarithmic transformation plus square root operation and then grayscale stretched to 256 levels to obtain a processed image.
  • the pre-processing operations described above may not be performed.
  • step S404 a series of binarization operations are performed after the initial image contrast is enhanced.
  • the binarization step be h, in the i-th binarization operation, the pixel point in the image whose pixel value p satisfies 0 ⁇ p ⁇ i*h(1 ⁇ i ⁇ N) is 255, and the remaining assignment value is 0. .
  • step S405 the connected regions of the binary images that meet the predetermined size are obtained and the upper and lower left and right boundaries of the connected regions are recorded, so that a lot of potential suspicious regions are obtained.
  • a feature generation and determination process is performed in which the feature information includes LBP features, shape features, and/or material features.
  • the feature information includes LBP features, shape features, and/or material features.
  • LBP features local texture features of the suspect region, such as LBP features
  • the generated local texture feature is sent to the classification model to obtain a class probability distribution. It is judged in step S410 whether the output result using the LBP feature index is suspect or not suspected based on the probability distribution.
  • step S406 all suspicious regions are traversed and contour shape features are extracted.
  • the domain uses an edge extraction feature extraction algorithm such as Canny to obtain the contour line of the suspect object in the suspicious area, and then calculates the centroid of the area enclosed by the contour line from the contour line, and uses the centroid as the coordinate origin to find the contour line and
  • the ( ⁇ , ⁇ ) curve between the origins of the coordinates. 7A, 7B, 7C, 7D, 7E, 7F, 7G, 7H, and 7I are ( ⁇ , ⁇ ) characteristic curves describing the use of a high atomic number material model in a scheme according to an embodiment of the present disclosure.
  • a schematic diagram in which the horizontal axis represents the angle ⁇ and the vertical axis represents the amplitude ⁇ .
  • the obtained ( ⁇ , ⁇ ) curve has a certain discontinuity, so the largest ⁇ jump in the ( ⁇ , ⁇ ) curve is recorded by ⁇ _jump.
  • ⁇ _jump Variable value.
  • the threshold is set to ⁇ /6 in this embodiment
  • the possibility that the rule object exists in the suspect area is very low.
  • the number of points on different contour lines is not the same, and the obtained ( ⁇ , ⁇ ) feature lengths are different. It is necessary to uniformly adjust the feature length to 200 dimensions for feature similarity comparison.
  • the contour of the object is simple and complex, and the fractal dimension of the boundary curve can be calculated. In general, the more complex the curve, the larger the fractal dimension, so the fractal dimension can also be used as a characteristic indicator to judge the shape and complexity of the curve shape.
  • the similarity between the shape of the hidden object and the standard shape in the suspicious area can be measured from three aspects of the Euclidean distance, the amplitude consistency, and the ⁇ hop value.
  • a certain contour S 0 be characterized by a 200-dimensional vector V 0
  • the standard contour vector is V st
  • ; the amplitude consistency is defined as The percentage P of the corresponding component in the same range of values.
  • L ⁇ 4 P ⁇ 0.9 and ⁇ max ⁇ ⁇ / 6, it is explained that the contour S 0 is similar to the standard contour.
  • step S411 the material image data is input, and in step S412, the proportion of the high atomic number pixel points in the suspect region is counted, and according to the given threshold value, it is possible to more accurately determine whether the current region contains the high atomic number substance. If the high atomic number detection process only receives fluoroscopic image data, the module can still give a judgment. Using the pre-trained classification model to calculate the shape characteristics of the suspicious area, it can be judged whether the fluoroscopic image contains high atomic number substances.
  • the comprehensive judgment refers to a step of comprehensively considering two or more determination results.
  • comprehensive judgment can be realized by methods such as decision fusion.
  • the method used in the present disclosure is to use the voting method to realize the judgment that if two or more input determination results are suspect, a warning is given and the suspicious area coordinates are output, otherwise the output is not suspect.
  • the high atomic number material detection result is marked with a red rectangular box or otherwise highlighted.
  • FIG. 5 is a schematic flow chart depicting a single energy mode in accordance with an embodiment of the present disclosure.
  • the embodiment of Fig. 5 reduces the input of the material image compared to the embodiment of Fig. 4, i.e., in the single energy mode, relying solely on the transmission image to determine whether or not the high atomic number material is contained in the inspected cargo. Therefore, the respective steps S501-S510 in FIG. 5 are the same as the steps S401-S410 shown in FIG. 4, respectively, in which the determination result of the shape feature and the determination result of the LBP feature are comprehensively determined in step S511, and the result is output, for example, the goods in step S512. Is it safe or a warning.
  • aspects of the embodiments disclosed herein may be implemented in an integrated circuit as a whole or in part, as one or more of one or more computers running on one or more computers.
  • a computer program eg, implemented as one or more programs running on one or more computer systems
  • implemented as one or more programs running on one or more processors eg, implemented as one or One or more programs running on a plurality of microprocessors, implemented as firmware, or substantially in any combination of the above, and those skilled in the art, in accordance with the present disclosure, will be provided with design circuitry and/or write software and / or firmware code capabilities.
  • signal bearing media include, but are not limited to, recordable media such as floppy disks, hard drives, compact disks (CDs), digital versatile disks (DVDs), digital tapes, computer memories, and the like; and transmission-type media such as digital and / or analog communication media (eg, fiber optic cable, waveguide, wired communication link, wireless communication link, etc.).

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Multimedia (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)
  • Image Processing (AREA)

Abstract

一种检查货物的方法和系统，该方法包括步骤：获得被检查货物的透射图像；对所述透射图像进行处理得到可疑区域；提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果；提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果；以及综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质。

Description

检查货物的方法、系统和装置 技术领域
本公开涉及辐射图像中的嫌疑物体的自动检测，具体而言就是涉及大型集装箱扫描系统中高原子序数物质的检查方法和系统。
背景技术
打击携带运输高原子序数物质如核材料、核武器和放射性散布装置具有重大的意义。自美国9·11事件以来，防范和处置核与辐射恐怖袭击已成为世界各国政府所面临的重大社会安全问题。与其它恐怖袭击相比，粗制滥造的核装置恐怖袭击发生的概率虽然相对比较低，但袭击的后果将会十分严重，不仅会导致人员的大量伤亡，而且会破坏经济发展，更会引发社会的恐慌与动荡。因此对用于制造核装置的核材料的监管已成为国际社会极为重视的核安全问题。
进入21世纪以来，每年累计有数以亿计的集装箱在全球各地流转。近些年，恐怖主义日益猖獗，核与辐射恐怖袭击阴云密布，恐怖分子可能会将大规模杀伤性武器或放射性散布装置藏匿在运送商品的集装箱中，将其送往目标袭击地。因此，需要检查货物中是否藏有核材料、核武器和放射性散布装置等，在降低安全风险的同时又保障了商业运输快速平稳地流通。
发明内容
鉴于现有技术中的一个或多个技术问题，本公开提出了一种检查货物的方法和系统。
在本公开的一个方面，提出了一种检查货物的方法，包括步骤：获得被检查货物的透射图像；对所述透射图像进行处理得到可疑区域；提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果；提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果；以及综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质。
在一些实施例中，所述方法还包括步骤：获取所述可疑区域的等效原子序数信息；其中所述综合步骤中综合所述分类结果、所述比较结果和等效原子序数信息，确定所述可疑区域中包含高原子序数物质。
在一些实施例中，对所述透射图像进行处理得到可疑区域的步骤包括：对所述透射图像进行对比度增强；对增强了对比度的透射图像进行多次二值化操作；求出二值化后的透射图像的连通区域并确定连通区域的边界。
在一些实施例中，提取所述可疑区域的轮廓线形状特征的步骤包括：使用边缘提取算法求得所述可疑区域的轮廓线，进而求出所述轮廓线的(θ，ρ)特征曲线。
在一些实施例中，将所述轮廓线形状特征与事先创建的标准模板进行比较的步骤包括：计算所述轮廓线的(θ，ρ)特征曲线与标准模板之间的距离和幅值一致性以及所述轮廓线的(θ，ρ)特征曲线的θ跳变值；将所述距离、所述幅值一致性和所述θ跳变值分别与阈值进行比较。
在一些实施例中，使用边缘特征提取算法获得所述可疑区域中嫌疑物体的轮廓线，再计算出轮廓线所围成区域的质心，然后以计算的质心为坐标原点，求出轮廓线与坐标原点间的(θ，ρ)特征曲线。
在一些实施例中，所述方法还包括步骤：在所述透视图像中突出显示被认为是高原子序数物质的区域。
在本发明的另一方面，提出了一种检查货物的系统，包括：X射线扫描设备，获得被检查货物的透射图像；数据处理装置，对所述透射图像进行处理得到可疑区域，提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果，提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果，以及综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质。
在本公开的又一方面，公开了一种检查货物的装置，包括：对被检查货物的透射图像进行处理得到可疑区域的装置；提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果的装置；提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果的装置；以及综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质的装置。
利用上述方案，能够自动检查货物中的高原子序数物质。例如，判断集装箱中是否藏匿高原子序数物质如核材料、核武器和放射性散布装置。这样可以减轻人工判图的工作量，提高工作效率，减少可疑物体区域的漏报，减少误报。
附图说明
为了更好的理解本公开，将根据以下附图对本公开的实施例进行描述：
图1A和图1B示出了根据本公开实施例的货物检查系统的结构示意图；
图2是描述根据本公开实施例的货物检查方法的示意性流程图；
图3是描述根据本公开实施例的方案中创建分类模型的过程的示意性流程图；
图4是描述根据本公开实施例的双能模式中使用创建的模型进行检查的方法的示意性流程图；
图5是描述根据本公开实施例的单能模式中的示意性流程图；
图6是描述根据本公开实施例的方案中使用的高原子序数物质的模型的示意图；以及
图7A、图7B、图7C、图7D、图7E、图7F、图7G、图7H和图7I是描述根据本公开实施例的方案中使用高原子序数物质模型的(θ，ρ)特征曲线的示意图。
附图没有对实施例的所有电路或结构进行显示。贯穿所有附图相同的附图标记表示相同或相似的部件或特征。
具体实施方式
下面将详细描述本公开的具体实施例，应当注意，这里描述的实施例只用于举例说明，并不用于限制本公开。在以下描述中，为了提供对本公开的透彻理解，阐述了大量特定细节。然而，对于本领域普通技术人员显而易见的是：不必采用这些特定细节来实行本公开。在其他实例中，为了避免混淆本公开，未具体描述公知的电路、材料或方法。
在整个说明书中，对“一个实施例”、“实施例”、“一个示例”或“示例”的提及意味着：结合该实施例或示例描述的特定特征、结构或特性被包含在本公开至少一个实施例中。因此，在整个说明书的各个地方出现的短语“在一个实施例中”、“在实施例中”、“一个示例”或“示例”不一定都指同一实施例或示例。此外，可以以任何适当的组合和/或子组合将特定的特征、结构或特性组合在一个或多个实施例或示例中。此外，本领域普通技术人员应当理解，在此提供的附图都是为了说明的目的，并且附图不一定是按比例绘制的。这里使用的术语“和/或”包括一个或多个相关列出的项目的任何和所有组合。
针对现有技术中无法准确快速判断货物中的高原子序数物质的问题，本公开 提出了基于被检查货物中可疑区域的局部纹理特征和/或轮廓线形状特征来判断可疑区域是否包含高原子序数物质。例如，通过(θ，ρ)特征描述可疑物体的形状特征，用LBP(Local Binary Pattern)特征描述可疑物体的纹理特征。在其他实施例(例如双能系统)中可以进一步使用有效原子序数信息来进一步判断。这样能够利用图像的纹理特征、可疑物体的边缘信息和/或材料信息，简化高原子序数物质检测过程，缩短计算时间，减少误报。
图1A和图1B是根据本公开一个实施方式的检查系统的结构示意图。图1A示出了检查系统的俯视示意图，图1B示出了检查系统的正视示意图。如图1A和图1B所示，射线源110产生X射线，经过准直器120准直后，对移动的集装箱卡车140进行安全检查，由探测器150接收穿透卡车的射线，在诸如计算机之类的数据处理装置160得到透射图像，并且对透射图像进行处理得到判断结果。
图2是描述根据本公开实施例的货物检查方法的示意性流程图。如图2所示，根据本公开的实施例，在步骤S21通过扫描得到集装箱卡车140的透射图像，然后在步骤S22通过数据处理装置160对透射图像进行处理，得到可疑区域。
在步骤S23，提取可疑区域的局部纹理特征，利用事先创建的模型对可疑区域的局部纹理特征进行分类，得到分类结果。然后在步骤S24，数据处理装置160提取可疑区域的轮廓线形状特征，并将轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果。最后在步骤S25，综合分类结果和比较结果，确定可疑区域中包含高原子序数物质。
图3是描述根据本公开实施例的方案中创建分类模型的过程的示意性流程图。在步骤S31，输入高原子序数物质图像，然后在步骤S32，提取图像的局部纹理特征，例如LBP特征。该LBP特征是一种用来描述图像局部纹理的算子，它具有旋转不变性和灰度不变性等优点。鉴于核材料、核武器、放射性散布装置及其屏蔽装置都具有规则的外形如立方体、球体或圆柱体等。在图像二值化求连通区域过程中可以获得大量的可疑区域，基本上所有高原子序数模拟物质都包含在其中的一小部分区域中。如图6所示，其中都是含有高原子序数物质的图像区域。将这些区域保存并进行标记分类，不含有高原子序数物质的区域作为一类，再将包含的按物体轮廓线型更细致地分为圆类、椭圆类、矩形类、平行四边形类、水平半圆类和竖直半圆类，共计7大类，然后提取LBP特征。在步骤S33基于提取的LBP特征进行训练，接下来在步骤S34获得分类模型。虽然在上述实施例中是将高原子序数物质分成7类，但是本领域的技术人员也可以相当将高原子序数物 质的形状分成其他数目的种类。
图4是描述根据本公开实施例的双能模式中使用创建的模型进行检查的方法的示意性流程图。图5是描述根据本公开实施例的单能模式中的示意性流程图。高原子序数自动检测的输入的既可以是透视图像也可以是透视图像和材料图(等效原子序数图)。当输入透视图像和材料图时，高原子序数检测按照图4所示流程进行。当仅输入透视图像时，则按照图5所示流程进行检测。图4和图5都包括输入图像(S401，S411；S501)、预处理(S402，S403；S502，S503)、生成可疑区域(S404，S405；S504，S505)、特征生成及判断(S407，S408，S409，S410；S507，S508，S509，S510)和输出(S414；S512)这些过程。下面分别结合图4和图5进行详细的描述。
在双能模式下，通过高能和低能图像能够得到被检查货物的等效原子序数信息，该信息可以与形状信息一起用来判断图像中是否包含高原子序数物质。在步骤S401，输入透视图像。
然后对输入的图像进行预处理。在步骤S402，检测图像中的集装箱区域，使得后续仅仅针对集装箱图像进行处理，减少了数据量，然后在步骤S403对图像进行对比度增强。例如，为了提高检测效率，可以在预处理阶段经过集装箱体检测算法后得到货物区域，用于后续高原子序数物质检测。为了提高对比度，可以对图像经过对数变换或对数变换加上开平方运算后再进行灰度拉伸到256级，即可得到处理后图像。本领域的技术人员可以认识到，在其他的实施例中，可以不进行上述的预处理操作。
在步骤S404，初始图像对比度增强后就进行一系列的二值化操作。设二值化步长为h，第i次二值化操作中将图像中像素值p满足0≤p≤i*h(1≤i≤N)的像素点赋值为255，剩余的赋值为0。然后在步骤S405求出这些二值图像中符合规定的尺寸大小的连通区域并记录下这些连通区域的上下左右边界，这样就得到很多潜在的可疑区域。
接下来进行特征生成及判断过程，该过程中特征信息包括LBP特征、形状特征和/或材料特征。例如，在步骤S407，在高原子序数物质检测过程中提取可疑区域的局部纹理特征，例如LBP特征。在步骤S408，将生成的局部纹理特征送入分类模型得到类别概率分布。在步骤S410根据概率分布判断使用LBP特征指标的输出结果是嫌疑还是无嫌疑。
在步骤S406，遍历所有可疑区域并提取轮廓线形状特征。对于每个可疑区 域，使用例如Canny之类的边缘提特征提取算法获得可疑区域中嫌疑物体的轮廓线，再由轮廓线计算出轮廓线所围成区域的质心，并以质心为坐标原点，求出轮廓线与坐标原点间的(θ，ρ)曲线。图7A、图7B、图7C、图7D、图7E、图7F、图7G、图7H和图7I是描述根据本公开实施例的方案中使用高原子序数物质模型的(θ，ρ)特征曲线的示意图，其中横轴表示角度θ，纵轴表示幅值ρ。
由于轮廓线是一些离散点构成的集合，再加之可疑区域的多样性，使得求得的(θ，ρ)曲线具有一定的间断性，故用θ_jump记录(θ，ρ)曲线中最大的θ跳变值。记轮廓线S＝{x1，x2，…，xN}且点是顺时针排列的，相邻两点相对于坐标原点的θ夹角为Δθ＝{θ1，θ2，…，θN}。其中，
Figure PCTCN2016097579-appb-000001
θmax用于衡量轮廓线的平滑性能和间断程度的，当其值大于给定的阈值θthresh(本实施例中该阈值设置为π/6)，说明可疑区域存在规则物体的可能性很低。而且不同的轮廓线上点的数量不尽相同，所得到的(θ，ρ)特征长度也就不同，需要将特征长度统一调整为200维，以便进行特征相似性对比。此外，物体轮廓线有简单与复杂之别，可以计算边界曲线的分维值。一般来说，曲线越复杂分维数就越大，因此分维数可以也作为判断曲线形状特征和复杂程度的一项特征指标。
为简化计算，在步骤S409，本实施例中可以从欧氏距离、幅值一致性和θ跳变值这三个方面来衡量可疑区域中隐藏物体形状和标准形状之间的相似度。例如，设某轮廓线S0特征化后是200维向量V0，而标准轮廓向量是Vst，记二者的欧氏距离为L＝|Vst-V0|；幅值一致性定义为对应分量处于同一取值范围的百分比P。在本实施例中，当L≤4，P≥0.9且θmax≤π/6时，说明该轮廓线S0与标准轮廓是相似的。
在步骤S411，输入材料图像数据，在步骤S412统计出可疑区域中高原子序数像素点所占的比例，根据给定阈值就能更加准确地判断当前区域是否含有高原子序数物质。如果高原子序数检测过程只接收到透视图像数据时，模块仍能给出判断。使用提前训练好的分类模型再计算出可疑区域的形状特征，就可以判断出透视图像中是否含有高原子序数物质。
在步骤S413，综合判断是指对两个或两个以上的判定结果进行综合考虑的步骤。一般地，可采用决策融合等方法实现综合判断。本公开中使用的方法是利用投票方法实现判断，即若两个或两个以上的输入判定结果为有嫌疑，则给出警告并输出可疑区域坐标，否则输出无嫌疑。在一些实施例中，针对高原子序数物质检测结果，用红色矩形框标记出来，或者其他方式突出显示出来。
图5是描述根据本公开实施例的单能模式中的示意性流程图。图5的实施例相比于图4的实施例减少了材料图像的输入，也就是在单能模式下仅仅依赖于透射图像来判断被检查货物中是否包含高原子序数物质。因此，图5中的各个步骤S501-S510分别与图4所示的步骤S401-S410相同，在步骤S511综合判断形状特征的判断结果和LBP特征的判断结果，在步骤S512输出结果，例如该货物是安全还是提出警告。
以上的详细描述通过使用示意图、流程图和/或示例，已经阐述了检查方法和系统的众多实施例。在这种示意图、流程图和/或示例包含一个或多个功能和/或操作的情况下，本领域技术人员应理解，这种示意图、流程图或示例中的每一功能和/或操作可以通过各种结构、硬件、软件、固件或实质上它们的任意组合来单独和/或共同实现。在一个实施例中，本公开的实施例所述主题的若干部分可以通过专用集成电路(ASIC)、现场可编程门阵列(FPGA)、数字信号处理器(DSP)、或其他集成格式来实现。然而，本领域技术人员应认识到，这里所公开的实施例的一些方面在整体上或部分地可以等同地实现在集成电路中，实现为在一台或多台计算机上运行的一个或多个计算机程序(例如，实现为在一台或多台计算机系统上运行的一个或多个程序)，实现为在一个或多个处理器上运行的一个或多个程序(例如，实现为在一个或多个微处理器上运行的一个或多个程序)，实现为固件，或者实质上实现为上述方式的任意组合，并且本领域技术人员根据本公开，将具备设计电路和/或写入软件和/或固件代码的能力。此外，本领域技术人员将认识到，本公开所述主题的机制能够作为多种形式的程序产品进行分发，并且无论实际用来执行分发的信号承载介质的具体类型如何，本公开所述主题的示例性实施例均适用。信号承载介质的示例包括但不限于：可记录型介质，如软盘、硬盘驱动器、紧致盘(CD)、数字通用盘(DVD)、数字磁带、计算机存储器等；以及传输型介质，如数字和/或模拟通信介质(例如，光纤光缆、波导、有线通信链路、无线通信链路等)。
虽然已参照几个典型实施例描述了本公开，但应当理解，所用的术语是说明和示例性、而非限制性的术语。由于本公开能够以多种形式具体实施而不脱离公开的精神或实质，所以应当理解，上述实施例不限于任何前述的细节，而应在随附权利要求所限定的精神和范围内广泛地解释，因此落入权利要求或其等效范围内的全部变化和改型都应为随附权利要求所涵盖。

Claims (10)

  1. 一种检查货物的方法，包括步骤：
    获得被检查货物的透射图像；
    对所述透射图像进行处理得到可疑区域；
    提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果；
    提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果；以及
    综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质。
  2. 如权利要求1所述的方法，还包括步骤：
    获取所述可疑区域的等效原子序数信息；
    其中所述综合步骤中综合所述分类结果、所述比较结果和等效原子序数信息，确定所述可疑区域中包含高原子序数物质。
  3. 如权利要求1所述的方法，其中对所述透射图像进行处理得到可疑区域的步骤包括：
    对所述透射图像进行对比度增强；
    对增强了对比度的透射图像进行多次二值化操作；
    求出二值化后的透射图像的连通区域并确定连通区域的边界。
  4. 如权利要求1所述的方法，其中提取所述可疑区域的轮廓线形状特征的步骤包括：
    使用边缘提取算法求得所述可疑区域的轮廓线，进而求出所述轮廓线的(θ，ρ)特征曲线。
  5. 如权利要求4所述的方法，其中将所述轮廓线形状特征与事先创建的标准模板进行比较的步骤包括：
    计算所述轮廓线的(θ，ρ)特征曲线与标准模板之间的距离和幅值一致性以及所述轮廓线的(θ，ρ)特征曲线的θ跳变值；
    将所述距离、所述幅值一致性和所述θ跳变值分别与阈值进行比较。
  6. 如权利要求5所述的方法，其中使用边缘特征提取算法获得所述可疑区域中嫌疑物体的轮廓线，再计算出轮廓线所围成区域的质心，然后以计算的质心 为坐标原点，求出轮廓线与坐标原点间的(θ，ρ)特征曲线。
  7. 如权利要求1所述的方法，还包括步骤：
    在所述透视图像中突出显示被认为是高原子序数物质的区域。
  8. 一种检查货物的系统，包括：
    X射线扫描设备，获得被检查货物的透射图像；
    数据处理装置，对所述透射图像进行处理得到可疑区域，提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果，提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果，以及综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质。
  9. 如权利要求8所述的系统，其中使用边缘提取算法求得所述可疑区域的轮廓线，进而求出所述轮廓线的(θ，ρ)特征曲线。
  10. 一种检查货物的装置，包括：
    对被检查货物的透射图像进行处理得到可疑区域的装置；
    提取所述可疑区域的局部纹理特征，并利用事先创建的模型对所述可疑区域的局部纹理特征进行分类，得到分类结果的装置；
    提取所述可疑区域的轮廓线形状特征，并将所述轮廓线形状特征与事先创建的标准模板进行比较，得到比较结果的装置；以及
    综合所述分类结果和所述比较结果，确定所述可疑区域中包含高原子序数物质的装置。
PCT/CN2016/097579 2015-12-16 2016-08-31 检查货物的方法、系统和装置 Ceased WO2017101514A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
SG11201709757VA SG11201709757VA (en) 2015-12-16 2016-08-31 Methods, systems, and apparatuses for inspecting goods
BR112017025624-0A BR112017025624B1 (pt) 2015-12-16 2016-08-31 Método, sistema e aparelho para inspecionar bens, e, mídia legível por computador não transitória

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510941150.7A CN106885813B (zh) 2015-12-16 2015-12-16 检查货物的方法、系统和装置
CN201510941150.7 2015-12-16

Publications (1)

Publication Number Publication Date
WO2017101514A1 true WO2017101514A1 (zh) 2017-06-22

Family

ID=57083141

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/097579 Ceased WO2017101514A1 (zh) 2015-12-16 2016-08-31 检查货物的方法、系统和装置

Country Status (5)

Country Link
US (1) US10134123B2 (zh)
EP (1) EP3182335A3 (zh)
CN (1) CN106885813B (zh)
SG (1) SG11201709757VA (zh)
WO (1) WO2017101514A1 (zh)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108956657A (zh) * 2018-08-23 2018-12-07 深圳码隆科技有限公司 一种安检过程中液体识别方法及其装置
CN110705385B (zh) * 2019-09-12 2022-05-03 深兰科技（上海）有限公司 一种障碍物角度的检测方法、装置、设备及介质
WO2021119946A1 (en) 2019-12-16 2021-06-24 Mekitec Oy Food inspection solution
CN112364903A (zh) * 2020-10-30 2021-02-12 盛视科技股份有限公司 基于x光机的物品分析及多维图像关联方法与系统
CN115452870B (zh) * 2022-09-30 2025-07-22 北京航星机器制造有限公司 一种爆炸物识别系统和方法
CN118657928B (zh) * 2024-07-23 2025-07-25 民航机场规划设计研究总院有限公司 基于人工智能的行李安检危险物品自动识别方法及系统
CN120431469A (zh) * 2025-07-04 2025-08-05 沈阳安丰电子工程有限公司 一种基于图像识别的智慧安检方法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5367552A (en) * 1991-10-03 1994-11-22 In Vision Technologies, Inc. Automatic concealed object detection system having a pre-scan stage
US6018562A (en) * 1995-11-13 2000-01-25 The United States Of America As Represented By The Secretary Of The Army Apparatus and method for automatic recognition of concealed objects using multiple energy computed tomography
US6069936A (en) * 1997-08-18 2000-05-30 Eg&G Astrophysics Material discrimination using single-energy x-ray imaging system
CN103744120A (zh) * 2013-12-30 2014-04-23 中云智慧（北京）科技有限公司 一种违禁品辅助鉴别方法及装置
CN104502381A (zh) * 2015-01-06 2015-04-08 安徽中福光电科技有限公司 一种解决通道式x光机实用性问题的“6331”方法
CN104750697A (zh) * 2013-12-27 2015-07-01 同方威视技术股份有限公司 基于透视图像内容的检索系统、检索方法以及安全检查设备

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7337686B2 (en) 2004-09-10 2008-03-04 Qylur Security Systems, Inc. Multi-threat detection system
US8600149B2 (en) 2008-08-25 2013-12-03 Telesecurity Sciences, Inc. Method and system for electronic inspection of baggage and cargo
EP2430396B1 (en) * 2009-05-16 2020-01-15 Rapiscan Systems, Inc. Systems and methods for automated, rapid detection of high-atomic-number materials
JP5685027B2 (ja) * 2010-09-07 2015-03-18 キヤノン株式会社 情報処理装置、物体把持システム、ロボットシステム、情報処理方法、物体把持方法およびプログラム
TW201344592A (zh) * 2012-04-27 2013-11-01 Altek Autotronics Corp 影像校正系統及其影像校正方法
US10297083B2 (en) * 2013-09-16 2019-05-21 Apple Inc. Method and system for determining a model of at least part of a real object
DE102014205447A1 (de) 2014-03-24 2015-09-24 Smiths Heimann Gmbh Detektion von Gegenständen in einem Objekt
WO2015150392A1 (en) * 2014-04-03 2015-10-08 Koninklijke Philips N.V. Examining device for processing and analyzing an image
US9633267B2 (en) * 2014-04-04 2017-04-25 Conduent Business Services, Llc Robust windshield detection via landmark localization
US10583354B2 (en) * 2014-06-06 2020-03-10 Lego A/S Interactive game apparatus and toy construction system
CN104181178B (zh) * 2014-08-18 2016-08-17 公安部第一研究所 一种通道式双视角x射线液态物品安全检查系统
CN105068134B (zh) * 2015-07-29 2017-09-15 公安部第一研究所 一种利用x射线多视角图像探测鞋中藏匿危险品的方法
US20170061686A1 (en) * 2015-08-28 2017-03-02 Hai Yu Stage view presentation method and system
WO2017040691A1 (en) * 2015-08-31 2017-03-09 Cape Analytics, Inc. Systems and methods for analyzing remote sensing imagery
US10625426B2 (en) * 2016-05-19 2020-04-21 Simbe Robotics, Inc. Method for automatically generating planograms of shelving structures within a store
US10650512B2 (en) * 2016-06-14 2020-05-12 The Regents Of The University Of Michigan Systems and methods for topographical characterization of medical image data

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5367552A (en) * 1991-10-03 1994-11-22 In Vision Technologies, Inc. Automatic concealed object detection system having a pre-scan stage
US6018562A (en) * 1995-11-13 2000-01-25 The United States Of America As Represented By The Secretary Of The Army Apparatus and method for automatic recognition of concealed objects using multiple energy computed tomography
US6069936A (en) * 1997-08-18 2000-05-30 Eg&G Astrophysics Material discrimination using single-energy x-ray imaging system
CN104750697A (zh) * 2013-12-27 2015-07-01 同方威视技术股份有限公司 基于透视图像内容的检索系统、检索方法以及安全检查设备
CN103744120A (zh) * 2013-12-30 2014-04-23 中云智慧（北京）科技有限公司 一种违禁品辅助鉴别方法及装置
CN104502381A (zh) * 2015-01-06 2015-04-08 安徽中福光电科技有限公司 一种解决通道式x光机实用性问题的“6331”方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CHEN, GONGYIN ET AL.: "Dual-Energy X-Ray Radiography for Automatic High-Z Material Detection", NUCLEAR INSTRUMENTS AND METHODS IN PHYSICS RESEARCH B, vol. 1-2, no. 261, 11 April 2007 (2007-04-11), pages 356 - 357, XP029159329 *
JITENDRA, M. ET AL.: "Contour and Texture Analysis for Image Segmentation", INTERNATIONAL JOURNAL OF COMPUTER VISION, vol. 43, no. 1, 1 January 2001 (2001-01-01), pages 7 - 27, XP009039552 *
PERTICONE, D. ET AL.: "Automatic Detection of High-Z Material in Cargo", SPIE-INT SOC OPTICAL ENGINEERING, 19 August 2010 (2010-08-19), pages 780511, XP055392704 *

Also Published As

Publication number Publication date
SG11201709757VA (en) 2017-12-28
EP3182335A3 (en) 2017-07-19
EP3182335A2 (en) 2017-06-21
US10134123B2 (en) 2018-11-20
US20170178312A1 (en) 2017-06-22
CN106885813B (zh) 2019-06-07
CN106885813A (zh) 2017-06-23
BR112017025624A2 (zh) 2018-08-07

Similar Documents

Publication Publication Date Title
WO2017101514A1 (zh) 检查货物的方法、系统和装置
US10509979B2 (en) Inspection methods and systems
CN108303747B (zh) 检查设备和检测枪支的方法
WO2015067208A1 (zh) 检查方法和设备
EP3349048A1 (en) Inspection devices and methods for detecting a firearm in a luggage
Jaccard et al. Tackling the X-ray cargo inspection challenge using machine learning
WO2019154383A1 (zh) 刀具检测方法及装置
JP6567703B2 (ja) 検査機器およびコンテナを検査する方法
Mekhalfa et al. Multiclass classification of weld defects in radiographic images based on support vector machines
Zhang et al. Label assignment matters: A gaussian assignment strategy for tiny object detection
Rogers et al. Detection of cargo container loads from X-ray images
Singh et al. Advancements in machine learning techniques for threat item detection in X-ray images: a comprehensive survey
US12580089B2 (en) Robust automatic tracking of individual triso-fueled pebbles through a novel application of x-ray imaging and machine learning
Li et al. Research on small target detection technology for river floating garbage based on improved yolov7
Jadhav et al. Enhanced Lung Cancer Detection and Classification Using YOLOv8
Saraswathi et al. Detection of juxtapleural nodules in lung cancer cases using an optimal critical point selection algorithm
Muthukkumarasamy et al. Intelligent illicit object detection system for enhanced aviation security
Muhtadan et al. Weld defect classification in radiographic film using statistical texture and support vector machine
CN111582367A (zh) 一种小金属威胁检测的方法
Liu et al. X-ray Security Contraband Detection Based on Improved YOLOX
ROUSSEL et al. Smart-RT: End-to-end automatic radiographic images analysis through a multi-stage object detection algorithm
BR112017025624B1 (pt) Método, sistema e aparelho para inspecionar bens, e, mídia legível por computador não transitória
WO2026051886A1 (zh) 级联的ct数据目标识别方法、装置及射线扫描检测系统
Moon et al. Extracting Breast Cancer Feature in Medical Images and Generating Its Parametric Pattern

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16874574

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 11201709757V

Country of ref document: SG

REG Reference to national code

Ref country code: BR

Ref legal event code: B01A

Ref document number: 112017025624

Country of ref document: BR

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 112017025624

Country of ref document: BR

Kind code of ref document: A2

Effective date: 20171129

122 Ep: pct application non-entry in european phase

Ref document number: 16874574

Country of ref document: EP

Kind code of ref document: A1