WO2016107478A1 - 车辆底盘检查方法和系统 - Google Patents

车辆底盘检查方法和系统 Download PDF

Info

Publication number
WO2016107478A1
WO2016107478A1 PCT/CN2015/098449 CN2015098449W WO2016107478A1 WO 2016107478 A1 WO2016107478 A1 WO 2016107478A1 CN 2015098449 W CN2015098449 W CN 2015098449W WO 2016107478 A1 WO2016107478 A1 WO 2016107478A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
chassis
image
chassis image
template
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/CN2015/098449
Other languages
English (en)
French (fr)
Inventor
李元景
李荐民
康克军
赵自然
刘耀红
李强
顾建平
胡峥
李营
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 MYPI2016703590A priority Critical patent/MY188116A/en
Publication of WO2016107478A1 publication Critical patent/WO2016107478A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis

Definitions

  • Embodiments of the present disclosure relate to vehicle chassis inspection, and more particularly to methods and systems for automatically inspecting vehicle chassis entrainment using image processing and pattern recognition techniques in the field of security inspection.
  • a vehicle chassis inspection method and system which are capable of automatically detecting entrainment in a chassis image.
  • a vehicle chassis inspection method comprising the steps of: acquiring a chassis image of a vehicle under inspection; acquiring a chassis image template of the same vehicle type as the vehicle from a database; and chassis of the inspected vehicle Aligning the image with the chassis image template; obtaining a difference between the registered chassis image and the registered chassis image template to obtain a variation region of the chassis image of the vehicle relative to the chassis image template; presenting the variation region .
  • the step of obtaining a chassis image template of the same vehicle type as the vehicle from the database comprises retrieving a chassis image template of the vehicle model from a database based on the unique identifier of the inspected vehicle.
  • the step of obtaining a chassis image template of the same vehicle type as the vehicle from the database comprises: obtaining a transmitted radiation image of the vehicle; extracting internal structural information of the vehicle from the transmitted radiation image, integrating the exterior of the vehicle The feature information retrieves the chassis image template of the model from the database.
  • the step of registering the chassis image of the inspected vehicle and the chassis image template comprises: rigidly registering a chassis image of the inspected vehicle and the chassis image template so as to The image is globally aligned; the chassis image of the inspected vehicle and the chassis image template are elastically registered to eliminate local distortion.
  • the step of rigid registration comprises: performing feature extraction on two images to obtain feature points; finding matching feature point pairs by performing similarity measure; obtaining image space coordinate transformation parameters by matching feature point pairs; and transforming parameters by coordinates Perform image registration.
  • the method further comprises the step of setting a pixel value whose absolute value in the variation region is less than a predetermined threshold to zero.
  • the method further includes the steps of: binarizing the image and performing a joint area analysis; and in a region where the area is smaller than or greater than a threshold, the pixel value is set to zero.
  • the method further comprises the step of setting a pixel value in a small area greater than zero or less than zero that occurs in pairs to zero.
  • a vehicle chassis inspection system comprising: a sensing device that acquires a chassis image of the inspected vehicle; and a data processing unit that acquires a chassis image template of the same vehicle type as the vehicle from the database, Registering a chassis image of the inspected vehicle and the chassis image template, and calculating a difference between the registered chassis image and the registered chassis image template to obtain a chassis image of the vehicle relative to the chassis image template a variable area; a display device that presents the changed area.
  • the vehicle chassis inspection system further includes: a radiation imaging system that obtains a transmitted radiation image of the vehicle; wherein the data processing unit extracts internal structural information of the vehicle from the transmitted radiation image, integrating the vehicle
  • the external feature information retrieves the chassis image template of the model from the database.
  • the above scheme can automatically detect hidden entrained objects from the chassis image through image processing and pattern recognition means, and has great practical value in the field of security inspection.
  • FIG. 1 shows a schematic diagram of a vehicle inspection system in accordance with an embodiment of the present disclosure:
  • FIG. 2 illustrates a flow chart of a vehicle type identification method in accordance with an embodiment of the present disclosure.
  • references to "one embodiment”, “an embodiment”, “an” or “an” or “an” or “an” or “an” In at least one embodiment.
  • the appearances of the phrase “in one embodiment”, “in the embodiment”, “the” Furthermore, the particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and/or sub-combination.
  • the term “and/or” as used herein includes any and all combinations of one or more of the associated listed items.
  • some embodiments of the present disclosure propose to perform an inspection using an automated method. For example, a chassis image of the inspected vehicle is acquired by a sensing device such as a CCD device, and then a chassis image template of the same vehicle type as the vehicle is acquired from the database. Further, the chassis image of the inspected vehicle and the chassis image template are registered, and then the registered chassis image and the registered chassis image template are deviated to obtain a change of the chassis image of the vehicle relative to the chassis image template. The area, which presents a varying area on the display, automatically indicates to the panelist the area where the entrainment may be present.
  • the safety inspection of the chassis can be automatically performed, and the efficiency and accuracy of the inspection are improved.
  • FIG. 1 shows a schematic diagram of a vehicle inspection system in accordance with an embodiment of the present disclosure.
  • an inspection system in accordance with an embodiment of the present disclosure involves automated inspection of a vehicle chassis using visible light images.
  • the system shown in FIG. 1 includes a sensing device 110, a radiation imaging system 150, a storage device 120, an image processing unit 140, and a display device 130.
  • sensing device 110 includes one or more sensors, such as CCD devices, etc., for obtaining chassis information for the vehicle, and the like.
  • the sensing device may include a camera for capturing a license plate image of the inspected vehicle; and an identification unit for identifying a license plate number of the inspected vehicle from the license plate image.
  • sensing device 110 includes a reader that reads an ID of the inspected vehicle from a radio frequency tag carried by the inspected vehicle.
  • the radiation imaging system 150 performs an X-ray scan of the inspected vehicle to obtain an X-ray image of the inspected vehicle.
  • the storage device 120 stores the X-ray image and the vehicle model database, including a chassis image template, a transmission image template, and the like.
  • the image processing unit 140 retrieves the vehicle model template corresponding to the vehicle from the vehicle model database, and determines a variation region between the obtained chassis image and the chassis template image. Display device 130 presents the changed region to the user.
  • the sensing device 110 obtains a chassis image of the vehicle.
  • the corresponding small vehicle can also be identified by the sensing device 110, generating a unique identification ID of the software system and the small vehicle, such as a license plate number.
  • the vehicle unique identification ID is a unique identifier for the small vehicle in the software system.
  • the identification ID may be data generated by the software system for the small vehicle, or may be identified by identifying the license plate number of the vehicle.
  • the current software system is identified by the license plate number.
  • the data processing unit 140 is responsible for performing a search for the template library to obtain a chassis template image corresponding to the small vehicle to be inspected.
  • the area of variation between the resulting chassis image and the chassis template image is determined.
  • Display device 130 presents the changed region to the user.
  • FIG. 2 shows a flow chart of a vehicle chassis inspection method in accordance with an embodiment of the present disclosure.
  • a chassis image of the inspected vehicle is acquired.
  • the chassis image acquisition uses a linear CCD camera.
  • This image is hereinafter referred to as a map to be detected.
  • image acquisition may include image correction and denoising. Since the image obtained by the linear camera is related to the vehicle speed, it is necessary to map each column of images with the time stamp acquired by the real-time speed measuring device (such as a radar) to obtain a uniform resolution image. Preferably, the physical size represented by each pixel in the set image is 5 mm * 5 mm. Image denoising can be achieved by a variety of algorithms. Preferably, it is implemented using a bilateral filter.
  • a chassis image template of the same vehicle type as the vehicle is acquired from the database.
  • This image is hereinafter referred to as a template map.
  • the template includes but not limited to: 1) using the license plate information to obtain the template image from the license plate duration image; 2) using the vehicle type recognition system to obtain the current vehicle chassis template image; 3) using the manual method , manually enter text, retrieve and select a template image.
  • the license plate recognition system is used, and the license plate is used as the identification information, and the historical database is searched for the image corresponding to the license plate under the same license plate as the template image.
  • the transmitted radiation image of the vehicle is obtained by the radiation inspection system 150, and then the data processing unit 140 extracts internal structural information of the vehicle from the transmitted radiation image, synthesizes the external characteristic information of the vehicle, and retrieves the vehicle model from the database. Chassis image template.
  • the "template image” is at least but not limited to an image.
  • multiple templates can be acquired, so that 1) can detect multiple changes of multiple templates, and multiple results can be integrated; 2) multiple templates can implement probabilistic templates; 3) partial stitching of multiple templates A preferred ideal template is implemented; 4) a better result is obtained by some means, such as selecting the template with the lowest noise as the preferred template map.
  • embodiments of the present disclosure are primarily based on the use of change detection rather than template optimization strategies.
  • step S23 the chassis image of the inspected vehicle and the chassis image template are registered; the image to be tested is registered with the template image.
  • This patent uses two sub-steps, rigid registration, and elastic registration to achieve optimized registration results. As a person skilled in the art, it is conceivable to use a plurality of registration algorithms to optimize the registration results, such as using gradient maps, transform domain features, multi-scale methods, etc. to improve the registration effect. The two images after registration have the same size.
  • Rigid registration is to globally align the image.
  • the flow is as follows: firstly, feature extraction is performed on the two images to obtain feature points; the matching feature point pairs are found by performing similarity measure; then the image space is obtained by matching feature point pairs. Coordinate transformation parameters; finally image registration by coordinate transformation parameters.
  • Feature extraction is the key in registration technology. Accurate feature extraction provides guarantee for the success of feature matching. Seeking feature extraction methods with good invariance and accuracy is crucial for matching accuracy.
  • SURF Speed Up Robust Features
  • the image to be tested is transmitted and deformed to the template image.
  • the deformation parameters were obtained using the Random Sample Consensus (RANSAC) algorithm.
  • the elastic registration of the image is primarily for accurate registration of the image to eliminate local distortion.
  • Elastic registration methods fall into two main categories: pixel-based methods and feature-based methods. After a lot of comparisons between calculation, validity, etc., the Demons elastic registration algorithm is preferably used to complete the process. However, other methods can be used by those skilled in the art for elastic registration.
  • step S24 the registered chassis image and the registered chassis image template are evaluated to obtain a variation region of the chassis image of the vehicle with respect to the chassis image template.
  • the registered image to be tested is subtracted from the registered template image to obtain a difference map.
  • the median method can be used to process the difference map. To find the median value of the difference map as a whole or part, the corresponding area minus this median value can better overcome this problem.
  • the variation area is presented.
  • the results indicate that there are many methods, which can be, but are not limited to, 1) direct colorization of the difference map; 2) binarization of the difference map, coloring the joint area; 3) binarization of the difference map, seeking connectivity The edge of the area, colored at the edges.
  • the colored difference map is merged with the image to be tested as the last The result is output.
  • the embodiment displays the results using a pseudo color method.
  • the absolute value of the difference map is obtained, and the range of values is stretched to, for example, [100, 255].
  • the image to be tested is converted into a black and white image with three channels of red, green and blue, and then the difference image is assigned to the red channel, so that the reddish effect of the entrained object is realized.
  • the difference map may be processed prior to presenting the varying region.
  • the goal of post-processing is to remove noise caused by stains, changes in vehicle position, and changes in the collection environment.
  • the difference map is also an image, the size is consistent with the image after registration, and the pixel is not zero, indicating that the image at the pixel changes.
  • the post-processing method can be, but is not limited to, 1) image denoising; 2) threshold processing, the absolute value is less than a threshold value is set to zero; 3) the image is binarized, and the communication area analysis is performed; the area is smaller or larger than In the region of a threshold value, the pixel value is set to zero; 4) Small regions larger than zero or less than zero, if present in pairs, are likely to be noise, and the pixel values of these regions are set to zero.
  • 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: Recording media, such as floppy disks, hard drives, compact disks (CDs), digital versatile disks (DVDs), digital tapes, computer memories, etc.; and transmission-type media, such as digital and/or analog communication media (eg, fiber optic cables, Waveguides, wired communication links, wireless communication links, etc.).

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)

Abstract

一种车辆底盘检查方法和系统。该方法包括步骤：获取被检查车辆的底盘图像；从数据库中获取与车辆相同车型的底盘图像模板；对被检查车辆的底盘图像和底盘图像模板进行配准；对配准后的底盘图像和配准后的底盘图像模板求差，得到车辆的底盘图像相对于底盘图像模板的变动区域；呈现所述变动区域。该方法通过图像处理与模式识别手段，从底盘图像中自动检测藏匿的夹带物，可应用于安检领域。

Description

车辆底盘检查方法和系统 技术领域
本公开的实施例涉及车辆底盘检查，具体而言，涉及到安全检查领域中，使用图像处理和模式识别技术自动检查车底盘夹带物的方法和系统。
背景技术
在汽车底盘非法藏匿夹带物，是犯罪分子瞒天过海的常用手段之一。对于这一巨大的安全隐患，业界有相当的重视，尤其是在机场方面，已经有相关标准或规范。目前相关的检测手段包括人工探视、底盘可见光图像检查等。人工探视方法的缺点显而易见，无法满足当前需求。而获取射线图像成本较高，且底盘图像较为杂乱，检测其中的夹带物也非常困难。可见光检查通常是通过CCD装置得到底盘的视频图像进行人工检查。这实际上仍旧是通过人工来检查，使得检查效率低，并且准确度不高。
发明内容
考虑到现有技术的上述问题，提出了一种车辆底盘检查方法和系统，能够自动检测底盘图像中的夹带物。
在本公开的一个方面，提出了一种车辆底盘检查方法，包括步骤：获取被检查车辆的底盘图像；从数据库中获取与所述车辆相同车型的底盘图像模板；对所述被检查车辆的底盘图像和所述底盘图像模板进行配准；对配准后的底盘图像和配准后的底盘图像模板求差，得到所述车辆的底盘图像相对于底盘图像模板的变动区域；呈现所述变动区域。
优选地，从数据库中获取与所述车辆相同车型的底盘图像模板的步骤包括：根据被检查车辆的唯一标识符从数据库中检索该车型的底盘图像模板。
优选地，从数据库中获取与所述车辆相同车型的底盘图像模板的步骤包括：获得所述车辆的透射辐射图像；从所述透射辐射图像中提取该车辆的内部结构信息，综合该车辆的外部特征信息，从数据库中检索该车型的底盘图像模板。
优选地，对所述被检查车辆的底盘图像和所述底盘图像模板进行配准的步骤包括：对所述被检查车辆的底盘图像和所述底盘图像模板进行刚性配准，以便对 图像进行全局变换对齐；对所述被检查车辆的底盘图像和所述底盘图像模板进行弹性配准，以便消除局部变形。
优选地，刚性配准的步骤包括：对两幅图像进行特征提取得到特征点；通过进行相似性度量找到匹配的特征点对；通过匹配的特征点对得到图像空间坐标变换参数；由坐标变换参数进行图像配准。
优选地，所述的方法还包括步骤：对所述变动区域内绝对值小于预定阈值的像素值设置为零。
优选地，所述的方法，还包括步骤：对图像进行二值化，且进行联通区域分析；面积小于或大于某门限的值的区域内，像素值置为零。
优选地，所述的方法，还包括步骤：将成对出现的大于零或小于零的小区域中的像素值置为零。
在本公开的另一方面，提出了一种车辆底盘检查系统，包括：传感设备，获取被检查车辆的底盘图像；数据处理单元，从数据库中获取与所述车辆相同车型的底盘图像模板，对所述被检查车辆的底盘图像和所述底盘图像模板进行配准，对配准后的底盘图像和配准后的底盘图像模板求差，得到所述车辆的底盘图像相对于底盘图像模板的变动区域；显示设备，呈现所述变动区域。
优选地，所述车辆底盘检查系统还包括：辐射成像系统，获得所述车辆的透射辐射图像；其中，所述数据处理单元从所述透射辐射图像中提取该车辆的内部结构信息，综合该车辆的外部特征信息，从数据库中检索该车型的底盘图像模板。
上述方案通过图像处理与模式识别手段，能够从底盘图像中自动检测藏匿的夹带物，在安检领域具有很大的实用价值。
附图说明
为了更好地理解本公开，将根据以下附图对本公开进行详细描述：
图1示出了根据本公开实施例的车辆检查系统的示意图：
图2示出了根据本公开实施例的车型识别方法的流程图。
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，未具体描述公知的结构、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，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。
在现有技术中，由于是通过人工来检查底盘图像，因此检查效率低，准确度不高。针对这个问题，本公开的一些实施例中提出了采用自动的方法进行检查。例如，通过诸如CCD器件之类的传感设备获取被检查车辆的底盘图像，然后从数据库中获取与该车辆相同车型的底盘图像模板。进而，对被检查车辆的底盘图像和底盘图像模板进行配准，接下来对配准后的底盘图像和配准后的底盘图像模板求差，得到该车辆的底盘图像相对于底盘图像模板的变动区域，在显示器上呈现变动区域，向判图人员自动指明可能存在夹带物的区域。通过上述方案，能够自动地对底盘进行安全检查，提高了检查的效率和准确度。
图1示出了根据本公开实施例的车辆检查系统的示意图。如图1所示，根据本公开实施例的检查系统涉及利用可见光图像对车辆底盘进行自动检查。
如图1所示的系统包括传感设备110、辐射成像系统150，存储设备120，图像处理单元140和显示设备130。
在一些实施例中，传感设备110包括一个或者多个传感器，例如CCD装置等，用来获得车辆的底盘信息等。在其他的实施例中，传感设备可以包括摄像机，用来捕获所述被检查车辆的车牌图像；和识别单元，用来从车牌图像识别所述被检查车辆的车牌号。在其他实施例中，传感设备110包括读取器，从所述被检查车辆所携带的射频标签读取所述被检查车辆的ID。
辐射成像系统150对被检查车辆进行X射线扫描，得到被检查车辆的X射线图像。存储设备120存储所述X射线图像以及车型模板数据库，包括底盘图像模板和透射图像模板等。
图像处理单元140从车型模板数据库中检索与该车辆相对应的车型模板，确定得到的底盘图像与底盘模板图像之间的变动区域。显示设备130向用户呈现所述变动区域。
例如，当有小型车辆需要检入时，传感设备110获得车辆的底盘图像。通过传感设备110也可以对相应小型车辆进行识别，生成软件系统与该小型车辆的唯一标识ID，如车牌号。该车辆唯一标识ID在该软件系统中是对该小型车辆过关的唯一标识。该标识ID可以软件系统针对该小型车辆生成的数据，也可以通过识别该车辆的车牌号，目前软件系统通过车牌号来标识。
例如，数据处理单元140负责针对模板库进行检索，得到和待检小型车辆相对应的底盘模板图像。确定得到的底盘图像和底盘模板图像之间的变动区域。显示设备130向用户呈现所述变动区域。
下面结合图2进一步说明根据本公开实施例的检查方法的流程图。图2示出了根据本公开实施例的车辆底盘检查方法的流程图。
在步骤S21，获取被检查车辆的底盘图像。优选的，底盘图像获取使用线性CCD摄像头。国内外已有较多相关产品和专利，此处不再赘述。此图像下文称为待检测图。
此外，图像的获取可以包括图像校正和去噪。由于线性照相机得到的图像和车速有关，所以需要将每一列图像与实时测速装置(如雷达)获取的时间戳对应，得到统一分辨率的图像。优选的，设定图像中每个像素代表的物理尺寸为5mm*5mm。图像去噪可以通过多种算法实现。优选的，使用双边滤波器实现。
在步骤S22，从数据库中获取与所述车辆相同车型的底盘图像模板。此图像下文称为模板图。模板的获取有多种手段，包括但不限于：1)利用车牌信息，从在本车牌历时图像中，获取模板图像；2)使用车型识别系统，获取当前车辆底盘模板图像；3)使用人工方式，人工输入文本，检索并选择模板图像。优选的，采用车牌识别系统，以车牌为标识信息，在历史数据库中搜索同车牌下，时间最近的该车牌对应的图像作为模板图像。
在其他实施例，通过辐射检查系统150获得车辆的透射辐射图像，然后数据处理单元140从透射辐射图像中提取该车辆的内部结构信息，综合该车辆的外部特征信息，从数据库中检索该车型的底盘图像模板。
本领域技术人员可以理解“模板图像”至少但不限于一幅图像。比如在使用车牌信息时，可以获取多幅模板，从而可以通过1)针对多个模板的变动检测，融合多个结果；2)多个模板可实现概率式模板；3)多个模板的局部拼接实现优选的理想模板；4)通过某种方式，如选择噪声最低的模板作为优选的模板图，从而得到更优的结果。为表达清楚，本公开的实施例以使用变动检测而不是模板优化策略为主。
在步骤S23，对所述被检查车辆的底盘图像和所述底盘图像模板进行配准；将待测图像与模板图像进行配准。本专利使用两个子步骤，即刚性配准，弹性配准实现优化的配准结果。作为本领域的技术人员，可想到到使用多种配准算法优化配准结果，比如使用梯度图、变换域特征、多尺度等方法提高配准效果。配准后的两幅图像具有相同的尺寸。
刚性配准是为了对图像进行全局变换对齐，其流程如下：首先对两幅图像进行特征提取得到特征点；通过进行相似性度量找到匹配的特征点对；然后通过匹配的特征点对得到图像空间坐标变换参数；最后由坐标变换参数进行图像配准。特征提取是配准技术中的关键，准确的特征提取为特征匹配的成功进行提供了保障。寻求具有良好不变性和准确性的特征提取方法，对于匹配精度至关重要。特征提取的方法很多。优选的，本实施例采用Speed Up Robust Features(SURF)提取图像的特征点与特征点处的描述子。之后，将待测图像对模板图像做透射变形。使用随机抽样一致(Random Sample Consensus，RANSAC)算法求取变形参数。
图像的弹性配准主要是为了对图像进行精确配准以消除局部变形。弹性配准方法主要分为两大类:基于像素的方法和基于特征的方法。经过计算量、有效性等多方面对比，优选地使用Demons弹性配准算法完成这一过程。但是本领域的技术人员可以使用其他的方法来进行弹性配准。
在步骤S24，对配准后的底盘图像和配准后的底盘图像模板求差，得到所述车辆的底盘图像相对于底盘图像模板的变动区域。优选的，配准后的待测图像减去配准后的模板图像得到差图。为减少采集环境变化带来的影响，可采用去中值法处理差图。对差图的整体或局部求中值，对应区域减去这个中值即可较好的克服这一问题。
在步骤S25，呈现所述变动区域。结果表示方法较多，手段可以但不限于是：1)直接对差图进行为彩色化；2)对差图二值化，对联通区域上色；3)对差图二值化，求联通区域边缘，在边缘处上色。彩色化的差图与待测图像融合，作为最后 的结果输出。
优选的，实施例使用伪彩色方法显示结果。首先求差图绝对值，并将值域范围拉伸到例如[100,255]。然后将待测图像转化为具有红、绿、蓝三通道的黑白图像，然后将差图赋给红色通道，即可实现夹带物红色显著显示效果。
在其他实施例中，在呈现变动区域之前，可以对差图进行处理。后处理的目标是，去掉由污渍、车辆位置变化、采集环境变化引起的噪声。差图也是一副图像，大小与配准后的图像一致，其像素不为零则表示该像素处图像有变动。后处理方法可以但不限于是：1)图像去噪；2)阈值处理，绝对值小于某门限的值置为零；3)对图像进行二值化，且进行联通区域分析；面积小于或大于某门限的值的区域内，像素值置为零；4)大于零或小于零的小区域如果成对出现，则很可能是噪声，这些区域的像素值置为零。
虽然通过上述步骤实现了本公开的一些实施例，从而能够检查车辆底盘。但是本领域的技术人员易于理解，每个步骤均可使用多种算法实现，而不局限于上述的具体步骤。
以上的详细描述通过使用示意图、流程图和/或示例，已经阐述了车辆底盘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. 一种车辆底盘检查方法，包括步骤：
    获取被检查车辆的底盘图像；
    从数据库中获取与所述车辆相同车型的底盘图像模板；
    对所述被检查车辆的底盘图像和所述底盘图像模板进行配准；
    对配准后的底盘图像和配准后的底盘图像模板求差，得到所述车辆的底盘图像相对于底盘图像模板的变动区域；
    呈现所述变动区域。
  2. 如权利要求1所述的方法，其中从数据库中获取与所述车辆相同车型的底盘图像模板的步骤包括：
    根据被检查车辆的唯一标识符从数据库中检索该车型的底盘图像模板。
  3. 如权利要求1所述的方法，其中从数据库中获取与所述车辆相同车型的底盘图像模板的步骤包括：
    获得所述车辆的透射辐射图像；
    从所述透射辐射图像中提取该车辆的内部结构信息，综合该车辆的外部特征信息，从数据库中检索该车型的底盘图像模板。
  4. 如权利要求1所述的方法，其中对所述被检查车辆的底盘图像和所述底盘图像模板进行配准的步骤包括：
    对所述被检查车辆的底盘图像和所述底盘图像模板进行刚性配准，以便对图像进行全局变换对齐；
    对所述被检查车辆的底盘图像和所述底盘图像模板进行弹性配准，以便消除局部变形。
  5. 如权利要求4所述的方法，其中刚性配准的步骤包括：
    对两幅图像进行特征提取得到特征点；
    通过进行相似性度量找到匹配的特征点对；
    通过匹配的特征点对得到图像空间坐标变换参数；
    由坐标变换参数进行图像配准。
  6. 如权利要求1所述的方法，还包括步骤：
    对所述变动区域内绝对值小于预定阈值的像素值设置为零。
  7. 如权利要求1所述的方法，还包括步骤：
    对图像进行二值化，且进行联通区域分析；
    面积小于或大于某门限的值的区域内，像素值置为零。
  8. 如权利要求1所述的方法，还包括步骤：
    将成对出现的大于零或小于零的小区域中的像素值置为零。
  9. 一种车辆底盘检查系统，包括：
    传感设备，获取被检查车辆的底盘图像；
    数据处理单元，从数据库中获取与所述车辆相同车型的底盘图像模板，对所述被检查车辆的底盘图像和所述底盘图像模板进行配准，对配准后的底盘图像和配准后的底盘图像模板求差，得到所述车辆的底盘图像相对于底盘图像模板的变动区域；
    显示设备，呈现所述变动区域。
  10. 如权利要求9所述的车辆底盘检查系统，还包括：辐射成像系统，获得所述车辆的透射辐射图像；
    其中，所述数据处理单元从所述透射辐射图像中提取该车辆的内部结构信息，综合该车辆的外部特征信息，从数据库中检索该车型的底盘图像模板。
PCT/CN2015/098449 2014-12-30 2015-12-23 车辆底盘检查方法和系统 Ceased WO2016107478A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
MYPI2016703590A MY188116A (en) 2014-12-30 2015-12-23 Methods and systems for inspecting vehicle chassis

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410841698.X 2014-12-30
CN201410841698.XA CN105807335B (zh) 2014-12-30 2014-12-30 车辆底盘检查方法和系统

Publications (1)

Publication Number Publication Date
WO2016107478A1 true WO2016107478A1 (zh) 2016-07-07

Family

ID=56284254

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/098449 Ceased WO2016107478A1 (zh) 2014-12-30 2015-12-23 车辆底盘检查方法和系统

Country Status (3)

Country Link
CN (1) CN105807335B (zh)
MY (1) MY188116A (zh)
WO (1) WO2016107478A1 (zh)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108460351A (zh) * 2018-02-28 2018-08-28 北京航星机器制造有限公司 一种便携式车底扫查系统及使用方法
CN110596776A (zh) * 2019-10-31 2019-12-20 厦门道塑汽车用品有限公司 一种汽车底盘安检检测装置
CN112488995A (zh) * 2020-11-18 2021-03-12 成都主导软件技术有限公司 列车自动化检修的智能判伤方法及系统
CN112991758A (zh) * 2021-03-24 2021-06-18 西安华旗电子技术有限公司 用于海关特殊监管区行政车辆货物夹带检查的抽查方法及装置
EP3726259A4 (en) * 2017-12-11 2021-08-25 Nuctech Company Limited VEHICLE CHASSIS SWEEPING SYSTEM AND SWEEPING METHOD
CN113822840A (zh) * 2020-06-19 2021-12-21 现代自动车株式会社 车辆车底的检查系统及方法
CN116109839A (zh) * 2023-02-15 2023-05-12 北京拙河科技有限公司 一种图片差异比对方法及装置
CN116468729A (zh) * 2023-06-20 2023-07-21 南昌江铃华翔汽车零部件有限公司 一种汽车底盘异物检测方法、系统及计算机

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108627348A (zh) * 2017-03-17 2018-10-09 北京爱德盛业科技有限公司 一种基于图像识别汽车底盘的检查方法
CN110363761A (zh) * 2019-07-22 2019-10-22 上海眼控科技股份有限公司 一种车辆底盘动态检测的起止标志检测系统及方法
CN117092713A (zh) 2020-05-28 2023-11-21 同方威视技术股份有限公司 建立车辆模板库的方法和系统
CN114801621A (zh) * 2022-05-17 2022-07-29 长春市华通机械电器有限公司 一种底盘结构件的控制臂系统
CN117152690A (zh) * 2023-08-30 2023-12-01 北京信路威科技股份有限公司 一种车辆底盘安检方法、装置和系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010191593A (ja) * 2009-02-17 2010-09-02 Honda Motor Co Ltd 対象物の位置検出装置及び位置検出方法
CN101945257A (zh) * 2010-08-27 2011-01-12 南京大学 基于监控视频内容提取车辆底盘图像的合成方法
CN102589458A (zh) * 2011-12-22 2012-07-18 上海一成汽车检测设备科技有限公司 汽车底盘钣金检测系统及其方法
CN103076641A (zh) * 2013-01-07 2013-05-01 河南科技大学 一种安全检测系统及检测方法
CN103338325A (zh) * 2013-06-14 2013-10-02 杭州普维光电技术有限公司 基于全景摄像机的车底盘图像采集方法
CN103646381A (zh) * 2013-11-22 2014-03-19 西安理工大学 一种线阵ccd的行进畸变校正方法
CN103984961A (zh) * 2014-05-30 2014-08-13 成都西物信安智能系统有限公司 一种用于检测车底异物的图像检测方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100588959C (zh) * 2006-10-10 2010-02-10 同方威视技术股份有限公司 基于辐射图像变动检测的小型车辆夹带物自动检测方法
US8792682B2 (en) * 2011-04-21 2014-07-29 Xerox Corporation Method and system for identifying a license plate
GB2512391B (en) * 2013-03-28 2020-08-12 Reeves Wireline Tech Ltd Improved borehole log data processing methods

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010191593A (ja) * 2009-02-17 2010-09-02 Honda Motor Co Ltd 対象物の位置検出装置及び位置検出方法
CN101945257A (zh) * 2010-08-27 2011-01-12 南京大学 基于监控视频内容提取车辆底盘图像的合成方法
CN102589458A (zh) * 2011-12-22 2012-07-18 上海一成汽车检测设备科技有限公司 汽车底盘钣金检测系统及其方法
CN103076641A (zh) * 2013-01-07 2013-05-01 河南科技大学 一种安全检测系统及检测方法
CN103338325A (zh) * 2013-06-14 2013-10-02 杭州普维光电技术有限公司 基于全景摄像机的车底盘图像采集方法
CN103646381A (zh) * 2013-11-22 2014-03-19 西安理工大学 一种线阵ccd的行进畸变校正方法
CN103984961A (zh) * 2014-05-30 2014-08-13 成都西物信安智能系统有限公司 一种用于检测车底异物的图像检测方法

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3726259A4 (en) * 2017-12-11 2021-08-25 Nuctech Company Limited VEHICLE CHASSIS SWEEPING SYSTEM AND SWEEPING METHOD
CN108460351A (zh) * 2018-02-28 2018-08-28 北京航星机器制造有限公司 一种便携式车底扫查系统及使用方法
CN110596776A (zh) * 2019-10-31 2019-12-20 厦门道塑汽车用品有限公司 一种汽车底盘安检检测装置
CN113822840A (zh) * 2020-06-19 2021-12-21 现代自动车株式会社 车辆车底的检查系统及方法
CN112488995A (zh) * 2020-11-18 2021-03-12 成都主导软件技术有限公司 列车自动化检修的智能判伤方法及系统
CN112488995B (zh) * 2020-11-18 2023-12-12 成都主导软件技术有限公司 列车自动化检修的智能判伤方法及系统
CN112991758A (zh) * 2021-03-24 2021-06-18 西安华旗电子技术有限公司 用于海关特殊监管区行政车辆货物夹带检查的抽查方法及装置
CN112991758B (zh) * 2021-03-24 2022-08-26 西安华旗电子技术有限公司 用于海关特殊监管区行政车辆货物夹带检查的抽查方法及装置
CN116109839A (zh) * 2023-02-15 2023-05-12 北京拙河科技有限公司 一种图片差异比对方法及装置
CN116468729A (zh) * 2023-06-20 2023-07-21 南昌江铃华翔汽车零部件有限公司 一种汽车底盘异物检测方法、系统及计算机
CN116468729B (zh) * 2023-06-20 2023-09-12 南昌江铃华翔汽车零部件有限公司 一种汽车底盘异物检测方法、系统及计算机

Also Published As

Publication number Publication date
CN105807335A (zh) 2016-07-27
MY188116A (en) 2021-11-21
CN105807335B (zh) 2019-12-03

Similar Documents

Publication Publication Date Title
CN105807335B (zh) 车辆底盘检查方法和系统
CN105809655B (zh) 车辆检查方法和系统
CN105574527B (zh) 一种基于局部特征学习的快速物体检测方法
US12417524B2 (en) Method and system of inspecting vehicle
CN102819740B (zh) 一种单帧红外图像弱小目标检测和定位方法
CN107909018B (zh) 一种稳健的多模态遥感影像匹配方法和系统
CN102789578B (zh) 基于多源目标特征支持的红外遥感图像变化检测方法
CN106384079A (zh) 一种基于rgb‑d信息的实时行人跟踪方法
WO2016034022A1 (zh) 车辆检查方法和系统
Ye et al. Fast and robust optical-to-SAR remote sensing image registration using region-aware phase descriptor
Wang et al. Building detection in high resolution satellite urban image using segmentation, corner detection combined with adaptive windowed hough transform
Shi et al. A method to detect earthquake-collapsed buildings from high-resolution satellite images
KR102235018B1 (ko) 해상도가 다른 영상을 정합하는 장치 및 방법
CN110084587B (zh) 一种基于边缘上下文的餐盘自动结算方法
CN119579990B (zh) 一种红外图像小目标识别方法、装置及可读存储介质
CN116664817B (zh) 基于图像差分的电力装置状态变化检测方法
Singla Technique of Image Registration in Digital Image processing-a review
KR20170131257A (ko) 머신 비전을 위한 영상 분석 방법 및 영상 분석 장치
Xue et al. Complete approach to automatic identification and subpixel center location for ellipse feature
Yang et al. A remote sensing imagery automatic feature registration method based on mean-shift
Zhao Comparative Study of Heterogeneous Remote Sensing Image Matching Algorithms
Zhang et al. Target Recognition System of Football Remote Sensing Image Based on Heapsort Algorithm
Jin et al. Improved moving target detection technology
Fan et al. A spatial feature enhanced MMI algorithm for multi-modal wild-fire image registration
CN118506306A (zh) 一种基于rgb和实体轮廓的二维地图语义定义方法和装置

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: 15875156

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: 15875156

Country of ref document: EP

Kind code of ref document: A1