WO2019028721A1 - 用于物品识别的方法、装置、设备和安检系统 - Google Patents

用于物品识别的方法、装置、设备和安检系统 Download PDF

Info

Publication number
WO2019028721A1
WO2019028721A1 PCT/CN2017/096748 CN2017096748W WO2019028721A1 WO 2019028721 A1 WO2019028721 A1 WO 2019028721A1 CN 2017096748 W CN2017096748 W CN 2017096748W WO 2019028721 A1 WO2019028721 A1 WO 2019028721A1
Authority
WO
WIPO (PCT)
Prior art keywords
line
item
article
segment
shape
Prior art date
Application number
PCT/CN2017/096748
Other languages
English (en)
French (fr)
Inventor
冷劲松
刘彦菊
李金嵘
Original Assignee
哈尔滨工业大学
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 哈尔滨工业大学 filed Critical 哈尔滨工业大学
Priority to PCT/CN2017/096748 priority Critical patent/WO2019028721A1/zh
Priority to CN201810062564.6A priority patent/CN109387880B/zh
Publication of WO2019028721A1 publication Critical patent/WO2019028721A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V5/00—Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity
    • G01V5/20—Detecting prohibited goods, e.g. weapons, explosives, hazardous substances, contraband or smuggled objects
    • G01V5/22—Active interrogation, i.e. by irradiating objects or goods using external radiation sources, e.g. using gamma rays or cosmic rays
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V5/00—Prospecting or detecting by the use of ionising radiation, e.g. of natural or induced radioactivity

Definitions

  • the present invention relates to the field of security inspection, and more particularly to a method, apparatus, device and security system for item identification.
  • Security systems are usually installed in subway stations, train stations, airports or other places with high security requirements to prevent criminals from using such dangerous goods as tools such as knives and firearms to enter these places for criminal activities.
  • the bags carried by people are required to be placed on the conveyor belt of the security inspection system.
  • the security inspection system uses X-rays to capture images of the items in the luggage, and finally the security inspection system or the security inspector identifies the shape of the objects in the image and based on the The shape is identified to determine if the item is a regulated dangerous item.
  • embodiments of the present invention provide a method, apparatus, apparatus, and security system for item identification that are capable of effectively identifying a regulated dangerous item in a deformed state.
  • a method for item identification includes: obtaining a feature line of the item using a three-dimensional image of the item, the feature line being a simple line of the outline of the item Representing; if the feature line indicates that the article is in a deformed state, acquiring a shape of an original state before deformation of the article based on the feature line; and displaying a shape of the original state of the article.
  • An apparatus for item identification includes: an extraction module for obtaining a feature line of the item using a three-dimensional image of the item, the feature line being a simple line of the outline of the item a processing module, configured to acquire a shape of an original state before deformation of the article based on the feature line if the feature line indicates that the article is in a deformed state; and a display module configured to display the The shape of the original state of the item.
  • An article identification device includes: a processor; and a memory recorded with executable instructions, wherein the executable instructions, when executed, cause the processor to perform the aforementioned method.
  • a security inspection system includes: a detecting device for taking a picture of an item placed on a conveyor belt for security inspection to obtain a three-dimensional image of the item; and a processing device connected to the detecting device And a display device for displaying the three-dimensional image and a shape of the original state.
  • the solution of the embodiment of the present invention acquires and displays the shape of the original state before deformation of the article in a deformed state, so that the shape of the original state of the article can be utilized to determine whether the article belongs to a controlled dangerous article. Therefore, the solution of the embodiment of the present invention can effectively identify the controlled dangerous goods in a deformed state as compared with the prior art.
  • FIG. 1 shows a block diagram of a security system in accordance with one embodiment of the present invention.
  • Fig. 2 shows a display example of a display device.
  • FIG. 3 shows a flow chart of a method for item identification in accordance with one embodiment of the present invention.
  • Figure 4 shows an exemplary diagram of a three-dimensional contour pattern, feature line and inversion of an item.
  • Figure 5 shows a flow chart of a method for item identification in accordance with one embodiment of the present invention.
  • Figure 6 shows a schematic diagram of an apparatus for item identification in accordance with one embodiment of the present invention.
  • Figure 7 shows a schematic diagram of an apparatus for item identification in accordance with one embodiment of the present invention.
  • Figure 8 shows a schematic diagram of a security system in accordance with one embodiment of the present invention.
  • FIG. 1 shows a block diagram of a security system in accordance with one embodiment of the present invention.
  • the security system 10 can include a detection device 20, a processing device 30, and a display device 40.
  • the detecting device 20, the processing device 30, and the display device 40 are connected to each other.
  • the detecting device 20 can take a picture of an item placed on the security belt to obtain a three-dimensional image of the item.
  • the detecting device 20 can take an image of an article by using X-rays or the like, so that a three-dimensional image of the article can be obtained even when the article is placed inside a bag or the like.
  • Processing device 30 may be a general purpose computer, an industrial computer, or other computing capable device.
  • the processing device 30 may perform image processing on the three-dimensional image captured by the detecting device 20 to acquire the shape of the original state before deformation of the article when the article is in a deformed state (eg, folded, bent, etc.), which will be described in detail below.
  • the display device 40 can display the three-dimensional image captured by the detecting device 20 and the shape of the original state before deformation of the article acquired by the processing device 30.
  • FIG. 2 shows a display example of the display device 40.
  • the display device 40 includes a main display portion on the left side and a sub display portion on the right side.
  • the main display portion is for displaying a three-dimensional image taken by the detecting device 20, for example, a three-dimensional image of the No. 1 item, a three-dimensional image of the No. 2 item, a three-dimensional image of the No. 3 item, and a three-dimensional image of the No. 4 item.
  • the sub display portion is for displaying the shape of the original state of the article acquired by the processing device 30, for example, the shape of the original state of the article No. 1, the shape of the original state of the article No. 2, the shape of the original state of the article No. 3, and the article No. 4 The shape of the original state.
  • FIG. 3 shows a flow chart of a method for item identification in accordance with one embodiment of the present invention.
  • the method 300 of FIG. 3 will be described below in conjunction with FIGS. 1 and 2.
  • the inspection device 20 takes a picture of the item OB placed on the security belt to capture a three-dimensional image of the item OB. .
  • the display device 40 displays a three-dimensional image of the item OB, for example, in its main display portion.
  • processing device 30 extracts a three-dimensional contour of item OB from the three-dimensional image of item OB.
  • FIG. 4 shows an example of a three-dimensional contour pattern ST1 extracted from a three-dimensional image of an article OB1.
  • the processing device 30 obtains a simple line representation of the three-dimensional contour of the item OB as a feature line for the item OB.
  • FIG. 4 shows an example of the feature line F1 of the article OB1.
  • processing device 30 determines if the feature line of item OB is not a straight line.
  • method 300 ends.
  • processing device 30 determines that article OB is in a deformed state (eg, folded, bent, etc.).
  • processing device 30 obtains a manner of inverting or expanding the feature lines of item OB into a straight line.
  • the manner in which the feature lines of the article OB are inverted or expanded into a line is the manner in which the feature lines of the article OB are converted into a plurality of line segments and the desired rotation angle of each of the plurality of line segments, wherein The desired rotation angle of each line segment indicates that the line segment is along a specified rotation direction SP (for example, clockwise or inverse) for a connection point between its line segment and its designated adjacent line segment in a straight line with respect to its designated adjacent line segment.
  • SP for example, clockwise or inverse
  • the hour hand direction the angle at which it needs to be rotated.
  • the approximation accuracy of approximating the feature line by using multiple segment lines can be adjusted by changing the length of the line segment. Generally, the smaller the length of the line segment, the higher the approximation accuracy.
  • the feature line F1 of the article OB1 is first converted into a plurality of line segments a1, b1, c1, d1, e1, f1, g1, h1, i1, j1, k1, l1. M1, n1 and o1. Then, starting from the end of the feature line F1 where the line segment a1 is located, the angle at which the line segment a1 needs to be rotated along the connection point between the line segment b1 and the line segment b1 in the specified rotation direction SP is calculated as an angle with respect to the adjacent line segment b1.
  • the line segment c1 is in a straight line around the connection point between the line segment b1 and the line segment c1 along the specified rotation direction SP as the required rotation angle AN2 of the line segment b1, and the calculation line segment c1 is in order to be adjacent to the adjacent line segment d1.
  • the calculation line segment n1 is in line with the adjacent line segment o1
  • the angle around which the connection point between the line segment n1 and the line segment o1 needs to be rotated along the specified rotation direction SP is taken as the required rotation angle AN14 of the line segment n1, and the required rotation angle AN15 of the line segment o1 is calculated (in this example, AN15) Equal to zero because line segment o1 is the last line segment and does not need to be rotated).
  • the line segment a1 first rotates the angle AN1 around the connection point between the line segment a1 and the line segment b1 along the specified rotation direction SP, then the line segments a1 and b1 follow the connection point between the line segment b1 and the line segment c1 along the specified rotation direction.
  • the SP rotation angle AN2 and then the line segments a1, b1, and c1 rotate the angle AN3, ... along the connection point between the line segment c1 and the line segment d1 in the specified rotation direction SP, and then the line segments a1, ..., m1, and n1 surround the line segment n1.
  • the connection point between the line segment o1 rotates the angle AN14 along the specified rotation direction SP, and the last line segments a1, ..., m1, n1, and o1 rotate the angle AN15 along the specified rotation direction SP, then the feature line F1 will be inverted or expanded.
  • a straight line FL the manner in which the characteristic line F1 of the article OB1 is inverted or expanded into a straight line is a manner in which the characteristic line portion F1 of the article OB1 is converted into a plurality of line segments a1, ..., o1 and a desired rotation of each of the line segments a1, ..., o1 Angles AN1, ..., AN15.
  • the processing device 30 inverts or expands the three-dimensional contour of the article OB in a manner that the feature line of the learned article OB is inverted or expanded into a line to obtain the shape of the original state of the article OB prior to deformation.
  • the three-dimensional contour pattern of the article OB is divided into a plurality of segmented portions in accordance with a manner in which the feature line of the article OB is converted into a plurality of line segments, wherein each segment portion and one of the plurality of segments Correspondingly, then, any one of the plurality of segment portions and all segment portions of the three-dimensional contour pattern located on a designated side of the any segment portion are located around the segment portion
  • the connecting line between the adjacent segment portions of the designated other side of the segment portion rotates the required rotation angle of the line segment corresponding to any of the segment portions along the specified rotation direction SP, thus obtaining the deformation of the OB.
  • the shape of the original original state is a manner in which the feature line of the article OB is converted into a plurality of line segments, wherein each segment portion and one of the plurality of segments Correspondingly, then, any one of the plurality of segment portions and all segment portions of the three-dimensional contour pattern located on a designated side of the any segment portion are located around the segment portion
  • the characteristic line F1 of the article OB1 is rotated.
  • the three-dimensional contour pattern ST1 of the article OB1 is divided into a plurality of segment portions a2, ..., o2 by means of a plurality of line segments a1, ..., o1, wherein the segment portions a2, ..., o2 are respectively associated with the line segments a1, ..., O1 corresponds.
  • the segmented portion a2 is first rotated around the connecting line between the segmented portion a2 and the adjacent segmented portion b2 on the other side of the segmented portion a2 in the specified rotational direction SP by the segmented portion.
  • the connecting line of the adjacent segment portion c2 rotates the required rotation angle AN2 of the line segment b1 corresponding to the segment portion b2 along the specified rotation direction SP, and then the segment portion c2 and the segment portion a2 on the side of the segment portion c2
  • b2 rotates the required rotation angle AN3 of the line segment c1 corresponding to the segment portion c2 around the connecting line of the segment portion c2 and the adjacent segment portion d2 on the other side of the segment portion c2 in the specified rotation direction SP.
  • the connecting line of the portion o2 rotates the line segment n1 corresponding to the segment portion n2 along the specified rotation direction SP.
  • the rotation angle AN14, and finally the segmentation portion o2 and the segment portions a2, b2, ..., n2 located on the side of the segment portion o2 are required to rotate the segment segment o1 corresponding to the segment portion o2 in the specified rotation direction SP.
  • the angle AN15 is rotated to obtain the shape Y1 of the original state before the deformation of the article OB1.
  • the processing device 30 displays the shape of the original state prior to the deformation of the resulting article OB on the display device 40, such as in the secondary display portion of the display device 40.
  • the processing device 30 may utilize the shape of the original state prior to the deformation of the resulting item OB to determine if the item OB belongs to a regulated dangerous item.
  • the processing device 30 may compare the shape of the original state before deformation of the article OB with a model of the pre-stored controlled dangerous goods to determine whether the article OB belongs to a regulated dangerous article. For another example, the processing device 30 may determine, for example, but not limited to, whether the article OB belongs to the tool by comparing the shape of the original state before the deformation of the article OB with the tool model stored in advance, and when the judgment result indicates that the article OB belongs to the tool. The tool nose angle of the article OB, and the article OB is determined to be a regulated dangerous article when the tool nose angle of the article OB satisfies the condition of the controlled tool.
  • the processing device 30 may indicate on the display device 40 that the item OB belongs to the regulated dangerous item. For example, as shown in FIG. 2, the processing device 30 displays a word "Control Tool!” on the display device 40 for items No. 2 and No. 3 that are determined to belong to the regulated dangerous goods.
  • the article in the deformed state is inverted or unfolded to obtain the shape of the original state before deformation of the article and it is judged whether or not the article belongs to the controlled danger based on the shape of the original state of the article. Therefore, even if the criminals use the shape memory material to make the controlled dangerous goods and deform them by folding or bending, etc., and the shape becomes different from the controlled dangerous goods, the solution of the embodiment can also be such a The controlled dangerous goods are identified, so that the solution of the embodiment can effectively identify the controlled dangerous goods in a deformed state.
  • method 300 includes blocks 318 and 320 to determine by the processing device 30 whether the item is a regulated dangerous item and to prompt when the determination is positive, however, The invention is not limited to this. In some other embodiments of the invention, method 300 may also include blocks 318 and 320, but rather the security inspector determines whether the article is a regulated hazard by observing the shape of the original state of item OB displayed on display device 40. Product.
  • the shape of the original state of the article is obtained by inverting or expanding the three-dimensional contour pattern of the article OB in such a manner that the feature line of the article is developed in a straight line, however, The invention is not limited to this. In some other embodiments of the present invention, the shape of the original state of the article may also be obtained by inverting or expanding the three-dimensional contour of the article OB by other forms.
  • Figure 5 shows a flow chart of a method for item identification in accordance with one embodiment of the present invention.
  • the method 500 shown in FIG. 5 can be performed, for example, by the processing device 30.
  • method 500 can include, at block 502, utilizing a three-dimensional image of the item to obtain A feature line of the item is obtained, the line of features being a simple line representation of the outline of the item.
  • the method 500 can also include, at block 504, if the feature line indicates that the item is in a deformed state, acquiring a shape of the original state prior to deformation of the item based on the feature line.
  • the method 500 can also include, at block 506, displaying the shape of the original state of the item.
  • the method 500 can also include determining whether the item belongs to a regulated dangerous item based on a shape of the original state; and, when determining that the item belongs to a regulated dangerous item, prompting the item to be subject to Control dangerous goods.
  • block 504 can include determining that the item is in a deformed state if the characteristic line is not a straight line, and expanding the feature line into a straight line if it is determined that the item is in the deformed state And expanding the three-dimensional contour of the item in a manner obtained to obtain a shape of the original state of the item, wherein the three-dimensional contour of the item is extracted from the image of.
  • the manner is a manner in which the characteristic line is converted into a plurality of line segments and a desired rotation angle of each of the plurality of line segments, wherein a desired rotation angle of each of the line segments indicates the line segment
  • An angle at which a connection point between the adjacent line segment and its specified adjacent line segment needs to be rotated in a straight line with respect to the specified adjacent line segment, and the three-dimensional contour pattern of the expanded object includes: Dividing the three-dimensional contour pattern into a plurality of segment portions according to the manner in which the feature line is converted into the plurality of line segments, wherein each of the plurality of segment portions corresponds to One of the plurality of line segments; and any segment of the plurality of segment portions and all segments of the three-dimensional contour pattern located on a designated side of the any segment portion Rotating the any portion of the segment along the specified direction of rotation about a connecting line between the segmented portion and the adjacent segmented portion on the designated other side of the segmented portion correspond Desired rotational angle of the line segment.
  • block 502 can include extracting a three-dimensional contour graphic of the item from the image; and obtaining a simple line representation of the three-dimensional contour graphic as the feature line.
  • the determining whether the item belongs to a controlled dangerous article comprises: if the shape of the original state indicates that the item is a tool, calculating a tool nose angle of the tool by using a shape of the original state And determining whether the tool belongs to the management based on the tool nose angle Dangerous goods.
  • Figure 6 shows a schematic diagram of an apparatus for item identification in accordance with one embodiment of the present invention.
  • the apparatus 600 shown in FIG. 6 can be implemented by means of software, hardware or a combination of software and hardware.
  • the apparatus 600 shown in FIG. 6 can be installed, for example, in the processing device 30.
  • device 600 can include an extraction module 602, a processing module 604, and a display module 606.
  • the extraction module 602 can be used to obtain a feature line of the item using a three-dimensional image of the item, the feature line being a simple line representation of the outline of the item.
  • the processing module 604 can be configured to acquire a shape of the original state of the article prior to deformation based on the feature line if the feature line indicates that the item is in a deformed state.
  • Display module 606 can be used to display the shape of the original state of the item.
  • the apparatus 600 may further include: a determining module configured to determine whether the item belongs to a regulated dangerous item based on a shape of the original state; and a prompting module configured to determine that the item belongs to a regulated In the case of dangerous goods, the item is indicated as a controlled dangerous goods.
  • the processing module 604 can include: means for determining that the item is in the deformed state if the feature line is not a straight line; for obtaining if the item is in the deformed state a module in which the feature lines are expanded in a straight line; and a module for expanding a three-dimensional contour pattern of the article in a manner acquired to obtain a shape of the original state of the article, wherein The three-dimensional contour of the item is extracted from the three-dimensional image.
  • the manner is a manner in which the characteristic line is converted into a plurality of line segments and a desired rotation angle of each of the plurality of line segments, wherein a desired rotation angle of each of the line segments indicates the The angle at which the line segment needs to be rotated in a straight line around its associated adjacent line segment in a straight line with respect to the specified adjacent line segment, and for expanding the manner in the acquired manner
  • the module of the three-dimensional contour of the article to obtain the shape of the original state of the article is further for: in accordance with the manner in which the feature line is converted into the plurality of line segments, the three-dimensional contour graphic Dividing into a plurality of segment portions, wherein each of the plurality of segment portions corresponds to one of the plurality of line segments; and, the plurality of segments Any of the segmented portions and all of the segmented portions of the three-dimensional contour pattern located on a designated side of the any of the segmented portions are around the any of the segmented portions and are located in any of The connecting line between
  • the extraction module 602 can include: a module for extracting a three-dimensional contour graphic of the article from the three-dimensional image; and a simple line representation for obtaining the three-dimensional contour graphic as the feature line Module.
  • Figure 7 shows a schematic diagram of an apparatus for item identification in accordance with one embodiment of the present invention.
  • the device 700 shown in FIG. 7 may be, for example, the processing device 30.
  • device 700 can include a processor 702 and a memory 704, wherein memory 704 records executable instructions, wherein the executable instructions, when executed, cause processor 702 to perform method 500.
  • FIG. 8 shows a schematic diagram of a security system in accordance with one embodiment of the present invention.
  • security system 800 can include detection device 802, processing device 804, and display device 806.
  • the detection device 802 can be used to take a picture of an item placed on the conveyor for security inspection to obtain a three-dimensional image of the item.
  • Processing device 804 is coupled to probing device 802 and can be used to perform method 500.
  • Display device 806 can be used to display the three-dimensional image and the shape of the original state.
  • Embodiments of the present invention provide a machine readable storage medium having recorded executable instructions, wherein the executable instructions, when executed, cause a machine to perform method 500.

Landscapes

  • Physics & Mathematics (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Geophysics (AREA)
  • Image Analysis (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Image Processing (AREA)

Abstract

一种用于物品识别的方法、装置、设备和安检系统，该方法包括：利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示；如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状；以及，显示所述物品的所述原始状态的形状。该方法、装置、设备和安检系统能够有效地识别处于变形状态的受管制危险品。

Description

用于物品识别的方法、装置、设备和安检系统 技术领域
本发明涉及安检领域，尤其涉及用于物品识别的方法、装置、设备和安检系统。
背景技术
在地铁站、火车站、机场或其它安全性要求较高的场所中通常设置有安检系统，以防止不法分子利用携带诸如刀具、枪支等受管制危险品进入这些场所进行犯罪活动。按照安检流程，人们携带的箱包被要求放置在安检系统的传输带上，然后安检系统利用X射线等拍摄箱包中的物品的图像，最后安检系统或安检员识别图像中的物品的形状并基于所识别的形状来判断物品是否属于受管制危险品。
随着科学技术的发展，出现了形状记忆聚合物及其复合材料、形状合金等新型形状记忆材料。不法分子可能会使用这类形状记忆材料来制作刀具、枪支等受管制危险品，并将其进行折叠或弯曲变形后在形状上变得不像刀具、枪支等受管制危险品。由于安检系统和安检员是基于物品的形状来判断其是否属于受管制危险品的，因此，安检系统和安检员很有可能不把这些折叠或弯曲变形的受管制危险品检测为受管制危险品，从而会造成受管制危险品漏检问题的发生，带来一定的安全隐患。
发明内容
考虑到现有技术的以上问题，本发明的实施例提供用于物品识别的方法、装置、设备和安检系统，其能够有效地识别处于变形状态的受管制危险品。
按照本发明的实施例的一种用于物品识别的方法，包括：利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条 表示；如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状；以及，显示所述物品的所述原始状态的形状。
按照本发明的实施例的一种用于物品识别的装置，包括：提取模块，用于利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示；处理模块，用于如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状；以及，显示模块，用于显示所述物品的所述原始状态的形状。
按照本发明的实施例的一种用于物品识别设备，包括：处理器；以及，存储器，其记录有可执行指令，其中，所述可执行指令当被执行时使得所述处理器执行前述的方法。
按照本发明的实施例的一种安检系统，包括：探测设备，用于对放置在传送带上进行安检的物品进行拍照，以得到所述物品的三维图像；处理设备，其与所述探测设备连接，用于执行前述的方法；以及，显示设备，用于显示所述三维图像和所述原始状态的形状。
从以上描述可以看出，本发明的实施例的方案获取和显示处于变形状态的物品的变形前的原始状态的形状，从而可以利用物品的原始状态的形状来判断物品是否属于受管制危险品，因此，与现有技术相比，本发明的实施例的方案能够有效地识别处于变形状态的受管制危险品。
附图说明
本发明的特征、特点和优点通过以下结合附图的详细描述将变得更加显而易见。
图1示出了按照本发明的一个实施例的安检系统的架构示意图。
图2示出了显示设备的一个显示例子。
图3示出了按照本发明的一个实施例的用于物品识别的方法的流程图。
图4示出了物品的三维轮廓图形、特征线和反演方式的示例图。
图5示出了按照本发明的一个实施例的用于物品识别的方法的流程图。
图6示出了按照本发明的一个实施例的用于物品识别的装置的示意图。
图7示出了按照本发明的一个实施例的用于物品识别的设备的示意图。
图8示出了按照本发明的一个实施例的安检系统的示意图。
具体实施方式
下面，详细描述本发明的各个实施例。
图1示出了按照本发明的一个实施例的安检系统的架构示意图。如图1所示，安检系统10可以包括检测设备20、处理设备30和显示设备40。检测设备20、处理设备30和显示设备40相互连接。
检测设备20可以对放置在安检带上的物品进行拍照，以得到物品的三维图像。检测设备20可以利用X射线等来对物品进行拍照，这样即使物品被放置在箱包等内部时也能得到物品的三维图像。
处理设备30可以是通用计算机、工业计算机或其它具有计算能力的设备。处理设备30可以对检测设备20拍摄的三维图像进行图像处理，以当物品处于变形状态(例如，折叠，弯曲等)时获取物品的变形前的原始状态的形状，这将在下面详细描述。
显示设备40可以显示检测设备20拍摄的三维图像和处理设备30获取的物品的变形前的原始状态的形状。图2示出了显示设备40的一个显示例子。如图2所示，显示设备40包括左侧的主显示部分和右侧的副显示部分。主显示部分用于显示检测设备20拍摄的三维图像，例如，①号物品的三维图像、②号物品的三维图像、③号物品的三维图像和④号物品的三维图像。副显示部分用于显示处理设备30获取的物品的原始状态的形状，例如，①号物品的原始状态的形状、②号物品的原始状态的形状、③号物品的原始状态的形状和④号物品的原始状态的形状。
图3示出了按照本发明的一个实施例的用于物品识别的方法的流程图。 下面将结合图1和图2描述图3的方法300。
在方框302，检测设备20对放置在安检带上的物品OB进行拍照，以拍摄得到物品OB的三维图像。。
在方框304，显示设备40例如在其主显示部分中显示物品OB的三维图像。
在方框306，处理设备30从物品OB的三维图像中，提取物品OB的三维轮廓图形。图4示出了从一个物品OB1的三维图像中提取的三维轮廓图形ST1的例子。
在方框308，处理设备30获取物品OB的三维轮廓图形的简单线条表示，作为物品OB的特征线。图4示出了物品OB1的特征线F1的例子。
在方框310，处理设备30判断物品OB的特征线是否不是直线。
如果方框310的判断结果为否定，即物品OB的特征线是直线，则方法300结束。
在方框312，如果方框310的判断结果为肯定，则处理设备30确定物品OB处于变形状态(例如，折叠，弯曲等)。
在方框314，处理设备30获取将物品OB的特征线反演或展开成直线的方式。在一个方面，将物品OB的特征线反演或展开成直线的方式是物品OB的特征线被转换成多个线段的方式和该多个线段中的每一个线段的所需旋转角，其中，每一个线段的所需旋转角指示该线段为了与其指定的相邻线段在一条直线上绕着其和该指定的相邻线段之间的连接点沿着指定旋转方向SP(例如顺时针方向或逆时针方向)需要旋转的角度。利用多个段线来近似特征线的近似精度可以通过改变线段的长度来调整，一般来说，线段的长度越小，那么近似精度越高。
例如，如图4所示，对于物品OB1的特征线F1，首先将特征线F1转换成多个线段a1、b1、c1、d1、e1、f1、g1、h1、i1、j1、k1、l1、m1、n1和o1。然后，从特征线F1中线段a1所在的一端开始，计算线段a1为了与相邻的线段b1在一条直线上绕着其和线段b1之间的连接点沿着指定旋转方向SP需要旋转的角度作为线段a1的所需旋转角度AN1，计算线段b1为了与相邻 的线段c1在一条直线上绕着线段b1和线段c1之间的连接点沿着指定旋转方向SP需要旋转的角度作为线段b1的所需旋转角度AN2，计算线段c1为了与相邻的线段d1在一条直线上绕着线段c1和线段d1之间的连接点沿着指定旋转方向SP需要旋转的角度作为线段c1的所需旋转角度AN3，…，计算线段n1为了与相邻的线段o1在一条直线上绕着线段n1和线段o1之间的连接点沿着指定旋转方向SP需要旋转的角度作为线段n1的所需旋转角度AN14，以及，计算线段o1的所需旋转角度AN15(在这个例子中AN15等于零，因为线段o1是最后一个线段，不需要旋转)。显然，如果首先线段a1绕着线段a1和线段b1之间的连接点沿着指定旋转方向SP旋转角度AN1，然后线段a1和b1绕着线段b1和线段c1之间的连接点沿着指定旋转方向SP旋转角度AN2，接着线段a1、b1和c1绕着线段c1和线段d1之间的连接点沿着指定旋转方向SP旋转角度AN3，…，接下来线段a1、…、m1和n1绕着线段n1和线段o1之间的连接点沿着指定旋转方向SP旋转角度AN14，以及，最后线段a1、…、m1、n1和o1沿着指定旋转方向SP旋转角度AN15，那么特征线F1将反演或展开成一条直线FL。这里，将物品OB1的特征线F1反演或展开成直线的方式是物品OB1的特征线分F1被转换成多个线段a1、…，o1的方式和线段a1、…，o1各自的所需旋转角度AN1、…、AN15。
在方框316，处理设备30按照所获知的物品OB的特征线反演或展开成直线的方式，反演或展开物品OB的三维轮廓图形，以得到物品OB的变形前的原始状态的形状。在一个方面，按照物品OB的特征线被转换成多个线段的方式，将物品OB的三维轮廓图形划分成多个分段部分，其中，每一个分段部分与该多个线段的其中一个线段对应，然后，将该多个分段部分中的任一分段部分和该三维轮廓图形中位于该任一分段部分的指定一侧的所有分段部分绕着该任一分段部分与位于该任一分段部分的指定另一侧的相邻分段部分之间的连接线沿着指定旋转方向SP旋转该任一分段部分对应的线段的所需旋转角度，这样就得到OB的变形前的原始状态的形状。
例如，如图4所示，在第一步骤中，按照将物品OB1的特征线F1被转 换成多个线段a1、…，o1的方式，将物品OB1的三维轮廓图形ST1分成多个分段部分a2、…，o2，其中，分段部分a2、…，o2分别与线段a1、…，o1对应。在第二步骤中，首先将分段部分a2绕着分段部分a2与位于分段部分a2的另一侧的相邻分段部分b2之间的连接线沿着指定旋转方向SP旋转分段部分a2所对应的线段a1的所需旋转角度AN1，然后将分段部分b2和位于分段部分b2一侧的分段部分a2绕着分段部分b2与位于分段部分b2的另一侧的相邻分段部分c2的连接线沿着指定旋转方向SP旋转分段部分b2所对应的线段b1的所需旋转角度AN2，接着将分段部分c2和位于分段部分c2一侧的分段部分a2、b2绕着分段部分c2与位于分段部分c2的另一侧的相邻分段部分d2的连接线沿着指定旋转方向SP旋转分段部分c2所对应的线段c1的所需旋转角度AN3，…，接下来将分段部分n2和位于分段部分n2一侧的分段部分a2、b2、…、m2绕着分段部分n2与位于分段部分n2的另一侧的相邻分段部分o2的连接线沿着指定旋转方向SP旋转分段部分n2所对应的线段n1的所需旋转角度AN14，以及，最后将分段部分o2和位于分段部分o2一侧的分段部分a2、b2、…、n2沿着指定旋转方向SP旋转分段部分o2所对应的线段o1的所需旋转角度AN15，从而得到物品OB1的变形前的原始状态的形状Y1。
在方框318，处理设备30将所得到的物品OB的变形前的原始状态的形状显示在显示设备40上，例如显示在显示设备40的副显示部分中。
在方框320，处理设备30可以利用所得到的物品OB的变形前的原始状态的形状，来确定物品OB是否属于受管制危险品。
例如，处理设备30可以将物品OB的变形前的原始状态的形状与预先存储的受管制危险品的模型进行比较，以确定物品OB是否属于受管制危险品。又例如，处理设备30可以例如但不局限于通过将物品OB的变形前的原始状态的形状与预先存储的刀具模型进行比较来判断物品OB是否属于刀具，当判断结果表明物品OB属于刀具时计算物品OB的刀尖角度，以及，当物品OB的刀尖角度满足受管制刀具的条件时确定物品OB属于受管制危险品。
如果在方框320中确定物品OB不属于受管制危险品，则方法300结束。
在方框322，如果在方框320中确定物品OB属于受管制危险品，则处理设备30可以在显示设备40上指示物品OB属于受管制危险品。例如，如图2所示，处理设备30在显示设备40上针对被确定属于受管制危险品的②号物品和③号物品显示“管制刀具！”词语。
从以上描述可以看出，在本实施例中，处于变形状态的物品被反演或展开以得到物品的变形前的原始状态的形状并基于物品的原始状态的形状来判断物品是否属于受管制危险品，因此，即使不法分子使用形状记忆材料来制作受管制危险品并将其进行折叠或弯曲等变形后在形状上变得不像受管制危险品，本实施例的方案也能将这样的受管制危险品识别出来，从而，本实施例的方案能够有效地识别处于变形状态的受管制危险品。
其它变型
本领域技术人员将理解，虽然在上面的实施例中，方法300包括方框318和320，以由处理设备30判断物品是否属于受管制危险品并当判断结果为肯定时进行提示，然而，本发明并不局限于此。在本发明的其它一些实施例中，方法300也可以不包括方框318和320，而是由安检员通过观察显示设备40上显示的物品OB的原始状态的形状来确定物品是否属于受管制危险品。
本领域技术人员将理解，虽然在上面的实施例中，通过按照将物品的特征线展开成直线的方式反演或展开物品OB的三维轮廓图形，来得到物品的原始状态的形状，然而，本发明并不局限于此。在本发明的其它一些实施例中，也可以通过其它的形式反演或展开物品OB的三维轮廓图形来得到物品的原始状态的形状。
图5示出了按照本发明的一个实施例的用于物品识别的方法的流程图。图5所示的方法500例如可以由处理设备30执行。
如图5所示，方法500可以包括，在方框502，利用物品的三维图像来获 得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示。
方法500还可以包括，在方框504，如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状。
方法500还可以包括，在方框506，显示所述物品的所述原始状态的形状。
在第一方面，方法500还可以包括：基于所述原始状态的形状，确定所述物品是否属于受管制危险品；以及，当确定所述物品属于受管制危险品时，提示所述物品是受管制危险品。
在第二方面，方框504可以包括：如果所述特征线不是直线，则确定所述物品处于变形状态；如果确定所述物品处于所述变形状态，则获取将所述特征线展开成直线的方式；以及，按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状，其中，所述物品的所述三维轮廓图形是从所述图像中提取的。
在第三方面，所述方式是所述特征线被转换成多个线段的方式和该多个线段中的每一个线段的所需旋转角度，其中，每一个线段的所需旋转角度指示该线段为了与其指定的相邻线段在一条直线上绕着其和该指定的相邻线段之间的连接点沿着指定旋转方向需要旋转的角度，以及，所述展开所述物品的三维轮廓图形包括：按照所述特征线被转换成所述多个线段的所述方式，将所述三维轮廓图形划分成多个分段部分，其中，所述多个分段部分中的每一个分段部分对应于所述多个线段的其中一个线段；以及，将所述多个分段部分中的任一分段部分和所述三维轮廓图形中位于所述任一分段部分的指定一侧的所有分段部分绕着所述任一分段部分与位于所述任一分段部分的指定另一侧的相邻分段部分之间的连接线沿着所述指定旋转方向旋转所述任一分段部分对应的线段的所述所需旋转角度。
在第四方面，方框502可以包括：从所述图像中提取所述物品的三维轮廓图形；以及，获得所述三维轮廓图形的简单线条表示，作为所述特征线。
在第五方面，所述确定所述物品是否属于受管制危险品包括：如果所述原始状态的形状表明所述物品是刀具，则利用所述原始状态的形状来计算所述刀具的刀尖角度；以及，基于所述刀尖角度，判断所述刀具是否属于受管 制危险品。
图6示出了按照本发明的一个实施例的用于物品识别的装置的示意图。图6所示的装置600可以利用软件、硬件或软硬件结合的方式来实现。图6所示的装置600可以例如安装在处理设备30中。
如图6所示，装置600可以包括提取模块602、处理模块604和显示模块606。提取模块602可以用于利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示。处理模块604可以用于如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状。显示模块606可以用于显示所述物品的所述原始状态的形状。
在第一方面，装置600还可以包括：确定模块，用于基于所述原始状态的形状，确定所述物品是否属于受管制危险品；以及，提示模块，用于当确定所述物品属于受管制危险品时，提示所述物品是受管制危险品。
在第二方面，处理模块604可以包括：用于如果所述特征线不是直线，则确定所述物品处于所述变形状态的模块；用于如果确定所述物品处于所述变形状态，则获取将所述特征线展开成直线的方式的模块；以及，用于按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状的模块，其中，所述物品的所述三维轮廓图形是从所述三维图像中提取的。
在第三方面，所述方式是所述特征线被转换成多个线段的方式和所述多个线段中的每一个线段的所需旋转角度，其中，每一个线段的所需旋转角度指示该线段为了与其指定的相邻线段在一条直线上绕着其和该指定的相邻线段之间的连接点沿着指定旋转方向需要旋转的角度，以及，用于按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状的所述模块进一步用于：按照所述特征线被转换成所述多个线段的所述方式，将所述三维轮廓图形划分成多个分段部分，其中，所述多个分段部分中的每一个分段部分对应于所述多个线段的其中一个线段；以及，将所述多 个分段部分中的任一分段部分和所述三维轮廓图形中位于所述任一分段部分的指定一侧的所有分段部分绕着所述任一分段部分与位于所述任一分段部分的指定另一侧的相邻分段部分之间的连接线沿着所述指定旋转方向旋转所述任一分段部分对应的线段的所述所需旋转角度。
在第四方面，提取模块602可以包括：用于从所述三维图像中提取所述物品的三维轮廓图形的模块；以及，用于获得所述三维轮廓图形的简单线条表示，作为所述特征线的模块。
图7示出了按照本发明的一个实施例的用于物品识别的设备的示意图。图7所示的设备700例如可以是处理设备30。如图7所示，设备700可以包括处理器702和存储器704，其中，存储器704记录有可执行指令，其中，所述可执行指令当被执行时使得处理器702执行方法500。
图8示出了按照本发明的一个实施例的安检系统的示意图。如图8所示，安检系统800可以包括探测设备802、处理设备804和显示设备806。探测设备802可以用于对放置在传送带上进行安检的物品进行拍照，以得到所述物品的三维图像。处理设备804与探测设备802连接，可以用于执行方法500。显示设备806可以用于显示所述三维图像和所述原始状态的形状。
本发明的实施例提供一种机器可读存储介质，其记录有可执行指令，其中，所述可执行指令当被执行时使得机器执行方法500。
本领域技术人员应当理解，以上公开的各个实施例可以在不偏离发明实质的情况下做出各种改变和变型，这些改变和变型都落入在本发明的保护范围之内，因此，本发明的保护范围将由所附的权利要求书来定义。

Claims (14)

  1. 一种用于物品识别的方法，包括：
    利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示；
    如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状；以及
    显示所述物品的所述原始状态的形状。
  2. 如权利要求1所述的方法，其中，还包括：
    基于所述原始状态的形状，确定所述物品是否属于受管制危险品；以及
    当确定所述物品属于受管制危险品时，提示所述物品是受管制危险品。
  3. 如权利要求1或2所述的方法，其中，所述获取所述物品的原始状态的形状包括：
    如果所述特征线不是直线，则确定所述物品处于所述变形状态；
    如果确定所述物品处于所述变形状态，则获取将所述特征线展开成直线的方式；以及
    按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状，其中，所述物品的所述三维轮廓图形是从所述图像中提取的。
  4. 如权利要求3所述的方法，其中
    所述方式是所述特征线被转换成多个线段的方式和该多个线段中的每一个线段的所需旋转角度，其中，每一个线段的所需旋转角度指示该线段为了与其指定的相邻线段在一条直线上绕着其和该指定的相邻线段之间的连接点沿着指定旋转方向需要旋转的角度，以及
    所述展开所述物品的三维轮廓图形包括：按照所述特征线被转换成所述 多个线段的所述方式，将所述三维轮廓图形划分成多个分段部分，其中，所述多个分段部分中的每一个分段部分对应于所述多个线段的其中一个线段；以及，将所述多个分段部分中的任一分段部分和所述三维轮廓图形中位于所述任一分段部分的指定一侧的所有分段部分绕着所述任一分段部分与位于所述任一分段部分的指定另一侧的相邻分段部分之间的连接线沿着所述指定旋转方向旋转所述任一分段部分对应的线段的所述所需旋转角度。
  5. 如权利要求1或2所述的方法，其中，所述获得所述物品的特征线包括：
    从所述图像中提取所述物品的三维轮廓图形；以及
    获得所述三维轮廓图形的简单线条表示，作为所述特征线。
  6. 如权利要求2所述的方法，其中，所述确定所述物品是否属于受管制危险品包括：
    如果所述原始状态的形状表明所述物品是刀具，则利用所述原始状态的形状来计算所述刀具的刀尖角度；以及
    基于所述刀尖角度，判断所述刀具是否属于受管制危险品。
  7. 一种用于物品识别的装置，包括：
    提取模块，用于利用物品的三维图像来获得所述物品的特征线，所述特征线是所述物品的轮廓的简单线条表示；
    处理模块，用于如果所述特征线表明所述物品处于变形状态，则基于所述特征线来获取所述物品的变形前的原始状态的形状；以及
    显示模块，用于显示所述物品的所述原始状态的形状。
  8. 如权利要求7所述的装置，其中，还包括：
    确定模块，用于基于所述原始状态的形状，确定所述物品是否属于受管制危险品；以及
    提示模块，用于当确定所述物品属于受管制危险品时，提示所述物品是受管制危险品
  9. 如权利要求7或8所述的装置，其中，所述处理模块包括：
    用于如果所述特征线不是直线，则确定所述物品处于所述变形状态的模块；
    用于如果确定所述物品处于所述变形状态，则获取将所述特征线展开成直线的方式的模块；以及
    用于按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状的模块，其中，所述物品的所述三维轮廓图形是从所述三维图像中提取的。
  10. 如权利要求9所述的装置，其中
    所述方式是所述特征线被转换成多个线段的方式和所述多个线段中的每一个线段的所需旋转角度，其中，每一个线段的所需旋转角度指示该线段为了与其指定的相邻线段在一条直线上绕着其和该指定的相邻线段之间的连接点沿着指定旋转方向需要旋转的角度，以及
    用于按照所获取的方式来展开所述物品的三维轮廓图形，以得到所述物品的所述原始状态的形状的所述模块进一步用于：按照所述特征线被转换成所述多个线段的所述方式，将所述三维轮廓图形划分成多个分段部分，其中，所述多个分段部分中的每一个分段部分对应于所述多个线段的其中一个线段；以及，将所述多个分段部分中的任一分段部分和所述三维轮廓图形中位于所述任一分段部分的指定一侧的所有分段部分绕着所述任一分段部分与位于所述任一分段部分的指定另一侧的相邻分段部分之间的连接线沿着所述指定旋转方向旋转所述任一分段部分对应的线段的所述所需旋转角度。
  11. 如权利要求7或8所述的装置，其中，所述提取模块包括：
    用于从所述三维图像中提取所述物品的三维轮廓图形的模块；以及
    用于获得所述三维轮廓图形的简单线条表示，作为所述特征线的模块。
  12. 一种用于物品识别的设备，包括：
    处理器；以及
    存储器，其记录有可执行指令，其中，所述可执行指令当被执行时使得所述处理器执行权利要求1-6中的任意一个所述的方法。
  13. 一种机器可读存储介质，其记录有可执行指令，其中，所述可执行指令当被执行时使得机器执行权利要求1-6中的任意一个所述的方法。
  14. 一种安检系统，包括：
    探测设备，用于对放置在传送带上进行安检的物品进行拍照，以得到所述物品的三维图像；
    处理设备，其与所述探测设备连接，用于执行权利要求1-5中的任意一个所述的方法；以及
    显示设备，用于显示所述三维图像和所述原始状态的形状。
PCT/CN2017/096748 2017-08-10 2017-08-10 用于物品识别的方法、装置、设备和安检系统 WO2019028721A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2017/096748 WO2019028721A1 (zh) 2017-08-10 2017-08-10 用于物品识别的方法、装置、设备和安检系统
CN201810062564.6A CN109387880B (zh) 2017-08-10 2018-01-23 用于物品识别的方法、装置、设备和安检系统

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2017/096748 WO2019028721A1 (zh) 2017-08-10 2017-08-10 用于物品识别的方法、装置、设备和安检系统

Publications (1)

Publication Number Publication Date
WO2019028721A1 true WO2019028721A1 (zh) 2019-02-14

Family

ID=65273061

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/096748 WO2019028721A1 (zh) 2017-08-10 2017-08-10 用于物品识别的方法、装置、设备和安检系统

Country Status (2)

Country Link
CN (1) CN109387880B (zh)
WO (1) WO2019028721A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113888394A (zh) * 2021-09-18 2022-01-04 阿里巴巴（中国）有限公司 图像变形方法、装置及电子设备
CN114998740A (zh) * 2022-06-13 2022-09-02 中国电子科技集团公司第五十四研究所 一种基于线段分布的机场线性特征提取方法
CN116503912A (zh) * 2023-06-25 2023-07-28 山东艾克斯智能科技有限公司 基于电子图囊的安检预警方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110321800A (zh) * 2019-06-05 2019-10-11 浙江大华技术股份有限公司 一种安检物品显示的方法及装置
CN111008676B (zh) * 2019-12-27 2022-07-12 哈尔滨工业大学 一种安检方法及安检系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708372A (zh) * 2012-02-29 2012-10-03 北京无线电计量测试研究所 一种隐匿物品的自动检测和识别方法
CN103903303A (zh) * 2012-12-27 2014-07-02 清华大学 三维模型创建方法和设备
CN103900503A (zh) * 2012-12-27 2014-07-02 清华大学 提取形状特征的方法、安全检查方法以及设备
CN103971084A (zh) * 2013-02-01 2014-08-06 同方威视技术股份有限公司 薄片和块状违禁品、爆炸物或毒品识别方法
CN104134071A (zh) * 2014-06-20 2014-11-05 北京工业大学 一种基于颜色描述的可变形部件模型物体检测方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103308538A (zh) * 2013-06-17 2013-09-18 开平市中铝实业有限公司 一种公交车安检系统
US9188696B2 (en) * 2013-06-25 2015-11-17 Analogic Corporation Mitigation of radiation leakage via entry port and/or exit port of radiation system
CN103839067A (zh) * 2014-03-18 2014-06-04 潘跃 物品信息获取方法、装置及系统
CN105784731B (zh) * 2014-06-25 2019-02-22 同方威视技术股份有限公司 一种定位三维ct图像中的目标的方法和安检系统
CN105574533B (zh) * 2015-12-15 2018-01-12 徐庆 一种图像特征提取方法和装置
CN106056599B (zh) * 2016-05-26 2019-04-30 四川大学 一种基于物体深度数据的物体识别算法及装置
CN106651841B (zh) * 2016-12-02 2020-10-16 北京航星机器制造有限公司 一种用于安检图像复杂度的分析方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102708372A (zh) * 2012-02-29 2012-10-03 北京无线电计量测试研究所 一种隐匿物品的自动检测和识别方法
CN103903303A (zh) * 2012-12-27 2014-07-02 清华大学 三维模型创建方法和设备
CN103900503A (zh) * 2012-12-27 2014-07-02 清华大学 提取形状特征的方法、安全检查方法以及设备
CN103971084A (zh) * 2013-02-01 2014-08-06 同方威视技术股份有限公司 薄片和块状违禁品、爆炸物或毒品识别方法
CN104134071A (zh) * 2014-06-20 2014-11-05 北京工业大学 一种基于颜色描述的可变形部件模型物体检测方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113888394A (zh) * 2021-09-18 2022-01-04 阿里巴巴（中国）有限公司 图像变形方法、装置及电子设备
CN114998740A (zh) * 2022-06-13 2022-09-02 中国电子科技集团公司第五十四研究所 一种基于线段分布的机场线性特征提取方法
CN116503912A (zh) * 2023-06-25 2023-07-28 山东艾克斯智能科技有限公司 基于电子图囊的安检预警方法
CN116503912B (zh) * 2023-06-25 2023-08-25 山东艾克斯智能科技有限公司 基于电子图囊的安检预警方法

Also Published As

Publication number Publication date
CN109387880B (zh) 2021-05-07
CN109387880A (zh) 2019-02-26

Similar Documents

Publication Publication Date Title
WO2019028721A1 (zh) 用于物品识别的方法、装置、设备和安检系统
JP7113657B2 (ja) 情報処理装置、情報処理方法、及びプログラム
US11210557B2 (en) Recognition apparatus, recognition method, and recording medium
US10438055B2 (en) Human facial detection and recognition system
EP3242269A1 (en) Image analysis method and apparatus, and computer readable device
US20150063627A1 (en) Methods and apparatus to identify components from images of the components
Hernández-Vela et al. BoVDW: Bag-of-Visual-and-Depth-Words for gesture recognition
US20150139494A1 (en) Slow change detection system
CN111738976A (zh) 信息处理装置、用于控制信息处理装置的方法和存储介质
EP3300025A1 (en) Image processing device and image processing method
TW201616450A (zh) 點雲畫筆選取系統及方法
JPWO2009001876A1 (ja) 特徴属性計算装置、特徴量抽出装置とパターン照合装置と方法及びプログラム
JP2021077039A5 (ja) 画像処理装置、画像処理方法、及びプログラム
JP5093591B2 (ja) 三次元位置姿勢計測方法および装置
EP4075388A1 (en) Computing device and method for generating realistic synthetic image data
Mi et al. A novel method for recognizing face with partial occlusion via sparse representation
CN111222452A (zh) 一种人脸匹配方法、装置、电子设备及可读存储介质
Božek et al. Flexible manufacturing system with automatic control of product quality
Sudhakar et al. Speeding-up SIFT based copy move forgery detection using level set approach
Gao et al. Vehicle bottom anomaly detection algorithm based on SIFT
Ye et al. An improved algorithm for Harris corner detection
US20110150339A1 (en) Apparatus, method, and program for image processing
JP7230722B2 (ja) 画像処理装置及び画像処理方法
JP5993100B2 (ja) 画像処理装置および特定図形検出方法
CN103745220A (zh) 一种图像仿射局部不变特征的获取方法及装置

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: 17921427

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17921427

Country of ref document: EP

Kind code of ref document: A1