WO2019024723A1 - 特征点匹配结果处理方法和装置 - Google Patents

特征点匹配结果处理方法和装置 Download PDF

Info

Publication number
WO2019024723A1
WO2019024723A1 PCT/CN2018/096866 CN2018096866W WO2019024723A1 WO 2019024723 A1 WO2019024723 A1 WO 2019024723A1 CN 2018096866 W CN2018096866 W CN 2018096866W WO 2019024723 A1 WO2019024723 A1 WO 2019024723A1
Authority
WO
WIPO (PCT)
Prior art keywords
intersection
feature point
pair
image
feature
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/CN2018/096866
Other languages
English (en)
French (fr)
Inventor
曹维雨
陈翔
赵大川
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.)
Goertek Inc
Original Assignee
Goertek Inc
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 Goertek Inc filed Critical Goertek Inc
Priority to US16/620,790 priority Critical patent/US11100361B1/en
Publication of WO2019024723A1 publication Critical patent/WO2019024723A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/70Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/757Matching configurations of points or features
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/2163Partitioning the feature space

Definitions

  • the invention relates to the technical field of image feature matching in computer vision, and in particular to a method and a device for processing feature point matching results.
  • image feature point matching methods mainly include methods for constructing matching descriptors based on local texture information, for example, SIFT, SURF, GLOH and other matching descriptors.
  • This method mainly uses feature local texture information for feature point matching, firstly features. The point determines the support area, then divides the support area into a series of fixed-shaped sub-areas, and finally represents the image texture information contained in each sub-area as a vector and constitutes a matching descriptor.
  • the present invention provides a method and apparatus for processing feature point matching results, which can quickly and easily verify the accuracy of feature point matching results.
  • the present invention provides a method for processing feature point matching results, including:
  • a matching feature point pair between the first image and the second image is determined by a feature point matching algorithm, wherein the first image and the second image are placed in opposite directions;
  • Whether or not the feature point pair is a correct matching result is determined according to whether the straight line between the pair of feature points passes the judgment area.
  • a line equation is made between each pair of feature points to determine the intersection between each line and other lines, including:
  • determining a valid intersection group according to the distance between each intersection and other intersections including:
  • the distance between each intersection point and other N-1 intersection points is calculated according to the coordinates of each intersection point
  • intersection point is an effective intersection point
  • determining a geometric center point of the effective intersection group includes:
  • the effective intersection groups formed by the K effective intersections are determined according to the K effective intersections. Geometric center point
  • the valid intersection group is determined among the N randomly selected intersection points until the K valid intersection points are determined in the valid intersection group, and K/N is greater than or equal to the second ratio threshold.
  • determining a determination area of the feature point pair according to the geometric center point including:
  • the determination area of the feature point pair is set according to a preset parameter, and the determination area includes a rectangular area, a circular area or an elliptical area, and the corresponding parameters include length, width, radius and/or diagonal length.
  • determining whether the feature point pair is a correct matching result according to whether the straight line between the pair of feature points passes the determination area includes:
  • the straight line between the pair of feature points does not pass the judgment area, it is determined that the feature point pair is an erroneous matching result, and the erroneous matching result is deleted.
  • determining a matching feature point pair between the first image and the second image by using a feature point matching algorithm including:
  • Matching feature point pairs corresponding to one-to-one correspondence between the first image and the second image are obtained according to the matching descriptor matching.
  • the present application further provides a feature point matching result processing apparatus, including:
  • a first determining module configured to determine, by using a feature point matching algorithm, a matching feature point pair between the first image and the second image, where the first image and the second image are placed in opposite directions;
  • a second determining module configured to perform a linear equation between each pair of feature points to determine an intersection between each straight line and other straight lines
  • a third determining module configured to determine a valid intersection group and a geometric center point of the effective intersection group according to a distance between each intersection point and other intersection points;
  • a fourth determining module configured to determine a determination area of the feature point pair according to the geometric center point
  • a fifth determining module configured to determine, according to whether the straight line between the pair of feature points passes the determination area, whether the feature point pair is a correct matching result.
  • the third determining module is specifically configured to:
  • the distance between each intersection point and other N-1 intersection points is calculated according to the coordinates of each intersection point
  • the effective intersection groups formed by the K effective intersections are determined according to the K effective intersections. Geometric center point
  • the valid intersection group is determined among the N randomly selected intersection points until the K valid intersection points are determined in the valid intersection group, and K/N is greater than or equal to the second ratio threshold.
  • the fourth determining module is specifically configured to:
  • the determination area of the feature point pair is set according to a preset parameter, and the determination area includes a rectangular area, a circular area or an elliptical area, and the corresponding parameters include length, width, radius and/or diagonal length;
  • the fifth determining module is specifically used to:
  • the first image and the second image are first placed in reverse, and the feature point matching algorithm can be used to determine the matching feature point pair between the first image and the second image by using the feature point matching algorithm; secondly, the mathematical expression is used in each A linear equation between feature point pairs can quickly determine the intersection between each line and other lines; and then determine the effective intersection group and the geometric center point of the effective intersection group according to the distance between each intersection and other intersections; The determination area of the feature point pair is determined according to the geometric center point, so that whether the feature point pair is a correct matching result can be determined according to whether the straight line between the feature point pairs passes the judgment area.
  • the existing feature point matching algorithm can only detect the correctness of matching feature points according to the naked eye after the matching feature points are completed.
  • the technical solution of the present invention can quickly determine the correctness of the matching result of the feature point by mathematical expression.
  • the present invention can determine a judgment area according to the center point of the effective intersection group, and can verify the verification of the matching feature point by judging the size of the area. Correctness, therefore, the matching feature points can be verified 100% correctly by the verification method of the present invention.
  • FIG. 1 is a schematic flowchart of a method for processing feature point matching results according to an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a line intersection point of a feature point pair according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a determination area according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of a feature point matching result processing apparatus according to an embodiment of the present invention.
  • FIG. 1 is a schematic flowchart of a method for processing feature point matching results according to an embodiment of the present invention, as shown in FIG.
  • first image and the second image are placed in the same screen, and then one of the first image and the second image (the first image or the second image) is rotated by 180 degrees.
  • Specific matches include:
  • Matching feature point pairs corresponding to one-to-one correspondence between the first image and the second image are obtained according to the matching descriptor matching.
  • the feature point matching algorithm is a local feature of the image, it has good invariance to translation, rotation, scale scaling, brightness variation, occlusion and noise, and maintains a certain degree of stability for visual changes and affine transformation. Fast and accurate matching in the massive feature database; even a small number of objects can generate a large number of SIFT feature vectors, which is relatively fast. Therefore, in the embodiment of the present invention, the feature point matching algorithm is run on the first image and the second image, wherein the feature point matching algorithm includes, for example, a matching descriptor sub-method such as SIFT, SURF, GLOH, etc., mainly by using feature local texture information.
  • a matching descriptor sub-method such as SIFT, SURF, GLOH, etc.
  • Feature point matching first determining the support area for the feature point, then dividing the support area into a series of fixed-shaped sub-areas, and finally representing the image texture information contained in each sub-area as a vector and composing a matching descriptor, according to the matching descriptor Matching results in a one-to-one correspondence between the first image and the second image, that is, feature point pairs.
  • the straight line equation can represent any straight line in the plane, the relationship between x and y can be quickly and accurately obtained, and the operation speed is relatively fast. Therefore, in the embodiment of the present invention, the straight line calculation between each feature point pair is performed. a straight line equation that determines the intersection between each line and the other line;
  • FIG. 2 is a schematic diagram of a line intersection point of a feature point pair according to an embodiment of the present invention.
  • a feature point (X1, Y1) in a first image and a feature point (X2, Y2) in a second image are assumed.
  • determining the effective intersection group according to the distance between each intersection and other intersections includes:
  • the distance between each intersection point and other N-1 intersection points is calculated according to the coordinates of each intersection point
  • intersection point is an effective intersection point
  • determining the geometric center point of the effective intersection group includes:
  • the effective intersection groups formed by the K effective intersections are determined according to the K effective intersections. Geometric center point
  • the valid intersection group is determined among the N randomly selected intersection points until the K valid intersection points are determined in the valid intersection group, and K/N is greater than or equal to the second ratio threshold.
  • a distance threshold a a first ratio threshold b% (usually can be set to 90% or 80%), and a second The proportional threshold c% (usually can be set to 70% or 60%).
  • the setting of the distance threshold a is inversely proportional to the accuracy requirement of the feature point matching result that needs to be implemented by the present invention. It is assumed that the higher the accuracy requirement of the feature point matching result required by the present invention, the smaller the value of the distance threshold a may be. The lower the accuracy requirement of the feature point matching result that the present invention needs to achieve, the larger the value of the distance threshold a can be.
  • the distance greater than the first ratio threshold b% is less than the distance threshold a, record the intersection point as the effective intersection point; if the line equation selected this time, the number of effective intersection points is proportional ( If the number of effective intersections/the number of intersections selected in this time is less than the threshold value c% of the second ratio, then a set of linear equations is re-selected to determine the effective intersection; if the number of effective intersections generated by the linear equation selected this time is greater than or equal to The second proportional threshold c% records all valid intersections to generate a valid intersection group and calculates the geometric center point of the effective intersection group.
  • the average value of each effective intersection coordinate may be determined as the coordinate of the geometric center point. For example, if the coordinates of the three effective intersections are (9, 11), (10, 10) and (11, 9), then the coordinates of the geometric center points of the three effective intersections are ((9+10+11)/ 3, (11+10+9)/3).
  • the determination area of the feature point pair is set according to a preset parameter centering on the geometric center point, and the determination area includes a rectangular area, a circular area, or an elliptical area, and the corresponding parameters include length, width, radius, and/or Diagonal length.
  • the parameters such as the length, the width, the radius, and/or the diagonal length may be determined according to the accuracy requirement of the feature point matching result that needs to be implemented by the present invention. It is assumed that the accuracy requirement of the feature point matching result required by the present invention is higher, and the length is longer.
  • the parameters such as width, radius and/or diagonal length may be smaller.
  • the rectangular determination area shown in FIG. 3 in order to achieve high-precision feature point matching results, the rectangular determination area shown in FIG. 3 can be set relatively small.
  • the feature point pair is a correct matching result; if the straight line between the feature point pairs does not pass the determination area, it is determined The feature point pair is not the correct match result.
  • FIG. 4 is a schematic diagram of determining a feature point matching result according to the prior art. As shown in FIG. 4, when the first image and the second image are placed in the same picture in the forward direction, if the feature point matching result is correct, the matching is performed.
  • the line formed by the pair of points should be parallel or gradually gradual, as shown in Figure 3. However, this parallel or a regular gradient is not easy to express mathematically, resulting in feature points. The result of the matching result is difficult or the judgment is incorrect.
  • the first image and the second image are placed in the same picture, and then one of the first image and the second image (the first image or the second image) is rotated. 180 degree.
  • all straight lines of correctly matched feature point pairs will intersect at an intersection point. Therefore, it is easy to calculate the intersection point between the two linear equations mathematically (such as the linear equation between feature point pairs).
  • the intersection of the lines of all correctly matched feature point pairs may not be an exact point, so a judgment area is determined, when the line between the feature point pairs Through the determination area, it is determined that the feature point pair is a correct matching result; when the straight line between the feature point pairs does not pass the determination area, it is determined that the feature point pair is not a correct matching result.
  • the first image and the second image are first placed in reverse, and the feature point matching algorithm can be used to determine the matching feature point pair between the first image and the second image by using the feature point matching algorithm; secondly, the mathematical expression is used in each A linear equation between feature point pairs can quickly determine the intersection between each line and other lines; and then determine the effective intersection group and the geometric center point of the effective intersection group according to the distance between each intersection and other intersections; The determination area of the feature point pair is determined according to the geometric center point, so that whether the feature point pair is a correct matching result can be determined according to whether the straight line between the feature point pairs passes the judgment area.
  • the existing feature point matching algorithm can only detect the correctness of matching feature points according to the naked eye after the matching feature points are completed.
  • the technical solution of the present invention can quickly determine the correctness of the matching result of the feature point by mathematical expression.
  • the present invention can determine a judgment area according to the center point of the effective intersection group, and can verify the verification of the matching feature point by judging the size of the area. Correctness, therefore, the matching feature points can be verified 100% correctly by the verification method of the present invention.
  • FIG. 5 is a schematic structural diagram of a feature point matching result processing apparatus according to an embodiment of the present invention. As shown in FIG. 5, the method includes:
  • a first determining module configured to determine, by using a feature point matching algorithm, a matching feature point pair between the first image and the second image, where the first image and the second image are placed in opposite directions;
  • a second determining module configured to perform a linear equation between each pair of feature points to determine an intersection between each straight line and other straight lines
  • a third determining module configured to determine a geometric intersection point of the effective intersection group and the effective intersection group according to the distance between each intersection point and other intersection points;
  • a fourth determining module configured to determine a determination area of the feature point pair according to the geometric center point
  • a fifth determining module configured to determine, according to whether the straight line between the pair of feature points passes the determination area, whether the feature point pair is a correct matching result.
  • the second determining module is specifically configured to:
  • the third determining module is specifically configured to:
  • the distance between each intersection point and other N-1 intersection points is calculated according to the coordinates of each intersection point
  • the effective intersection groups formed by the K effective intersections are determined according to the K effective intersections. Geometric center point
  • the valid intersection group is determined among the N randomly selected intersection points until the K valid intersection points are determined in the valid intersection group, and K/N is greater than or equal to the second ratio threshold.
  • the fourth determining module is specifically configured to:
  • the determination area of the feature point pair is set according to a preset parameter, and the determination area includes a rectangular area, a circular area or an elliptical area, and the corresponding parameters include length, width, radius and/or diagonal length.
  • the fifth determining module is specifically configured to:
  • the straight line between the pair of feature points does not pass the judgment area, it is determined that the feature point pair is an erroneous matching result, and the erroneous matching result is deleted.
  • the device in the embodiment of the present invention may be the method shown in the foregoing embodiment, and the implementation principle and technical effects thereof are not described again.
  • the structure of the feature point matching result processing device includes a processor and a memory, and the memory is configured to store the feature point matching result processing device in the embodiment shown in FIG. And a processor configured to execute a program stored in the memory.
  • the program includes one or more computer instructions, wherein the one or more computer instructions are for execution by the processor.
  • the embodiment of the present invention further provides a computer storage medium, which is used for storing computer software instructions used in the method for processing feature point matching results in the embodiment shown in FIG. 1, wherein the computer software instructions are included for The program involved in the feature point matching result processing method in the embodiment shown in FIG. 1 is executed.
  • the embodiment of the invention further provides a computer storage medium for storing computer software instructions used by the feature point matching result processing device, wherein the computer software instruction comprises a feature point matching result for performing the above feature point matching result processing method The program involved in the processing device.
  • embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media includes both permanent and non-persistent, removable and non-removable media.
  • Information storage can be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include temporary storage of computer readable media, such as modulated data signals and carrier waves.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Multimedia (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Image Analysis (AREA)

Abstract

本发明公开了特征点匹配结果处理方法和装置,首先将第一图像和第二图像反向放置,利用特征点匹配算法确定第一图像和第二图像之间匹配的特征点对;其次利用数学表达方式,在每一特征点对之间做直线,确定每条直线与其他直线之间的交点;根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;最后根据几何中心点确定特征点对的判断区域,从而可以根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。因此,本发明的技术方案很容易通过数学表达方式判断特征点对匹配结果的正确性。

Description

特征点匹配结果处理方法和装置
交叉引用
本申请引用于2017年08月01日递交的名称为“一种特征点匹配结果处理方法和装置”的第201710648107.0号中国专利申请,其通过引用被全部并入本申请。
技术领域
本发明涉及计算机视觉中的图像特征匹配技术领域,尤其涉及一种特征点匹配结果处理方法和装置。
背景技术
特征匹配技术在图像检索、物体识别、视频跟踪以及增强现实等领域有重要应用。目前已有图像特征点匹配方法主要有基于局部纹理信息构造匹配描述子的方法,例如,SIFT、SURF、GLOH等匹配描述子,该类方法主要利用特征局部纹理信息进行特征点匹配,首先为特征点确定支撑区域,然后将支撑区域划分为一系列固定形状的子区域,最后将每个子区域内包含的图像纹理信息表示为向量并组成匹配描述子。
该方法由于简单且易于实现,在实际应用中使用较多。但是,该类法均存在共性的误差问题:为实现尺度不变性而进行尺度选择产生尺度误差;为实现旋转不变性而估计主方向产生主方向误差;固定位置的子区域划分产生边界误差。三种误差的存在影响了描述子的分辨能力,特别是在图像中存在形变的情况下匹配效果不好。
因此,迫切需要一种方法可以进一步验证特征点匹配结果的正确性。
发明内容
为了解决上述问题,本发明提供一种特征点匹配结果处理方法和装置,可以快速简便验证特征点匹配结果的准确性。
本发明提供一种特征点匹配结果处理方法,包括:
通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,其中,第一图像和第二图像反向放置;
在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点;
根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;
根据几何中心点确定特征点对的判断区域;
根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。
可选地,在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点,包括:
在每一特征点对之间做直线,计算每条直线的直线方程;
根据每条直线方程,确定每条直线方程之间的交点。
可选地,根据每个交点与其他交点之间的距离,确定有效交点组,包括:
在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
在得到的N-1个交点之间的距离中若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点;
将N个交点中的全部有效交点形成有效交点组。
可选地,确定有效交点组的几何中心点,包括:
若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形 成的有效交点组的几何中心点;
若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
可选地,根据几何中心点确定特征点对的判断区域,包括:
以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度。
可选地,根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果,包括:
若特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;
若特征点对之间的直线不通过所述判断区域,则确定该特征点对是错误的匹配结果,删除错误的匹配结果。
可选地,通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,包括:
分别确定第一图像和第二图像的特征点支撑区域;
将特征点支撑区域划分为多个固定形状的子区域;
将每个子区域内包含的图像纹理信息表示为向量并组成匹配描述子;
根据匹配描述子匹配得到第一图像和第二图像中一一对应的匹配的特征点对。
本申请还一种特征点匹配结果处理装置,其特征在于,包括:
第一确定模块,用于通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,其中,第一图像和第二图像反向放置;
第二确定模块,用于在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点;
第三确定模块,用于根据每个交点与其他交点之间的距离,确定有效交 点组以及有效交点组的几何中心点;
第四确定模块,用于根据几何中心点确定特征点对的判断区域;
第五确定模块,用于根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。
可选地,第三确定模块具体用于:
在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
在得到的N-1个交点之间的距离中,若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点,将N个交点中的全部有效交点形成有效交点组;
若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形成的有效交点组的几何中心点;
若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
可选地,第四确定模块具体用于:
以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度;
第五确定模块具体用于:
若特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;
若特征点对之间的直线不通过所述判断区域,则确定该特征点对不是正确的匹配结果,删除错误的匹配结果。
本发明实施例首先将第一图像和第二图像反向放置,利用特征点匹配算法可以快色确定第一图像和第二图像之间匹配的特征点对;其次利用数学表 达方式,在每一特征点对之间做直线方程,可以快速确定每条直线与其他直线之间的交点;进而根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;最后根据几何中心点确定特征点对的判断区域,从而可以根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。解决了现有的特征点匹配算法在匹配特征点完成后,只能根据肉眼检测匹配特征点的正确性而存在误差较大的问题。而本发明的技术方案通过数学表达方式可以快速地判断特征点对匹配结果的正确性,本发明可以根据有效交点组的中心点确定一个判断区域,通过判断区域的大小可以调整匹配特征点的验证正确性,因此,通过本发明的验证方法可以对匹配特征点进行百分之百正确性的验证。
附图说明
此处所说明的附图用来提供对本申请的进一步理解,构成本申请的一部分,本申请的示意性实施例及其说明用于解释本申请,并不构成对本申请的不当限定。在附图中:
图1为本发明一实施例提供的特征点匹配结果处理方法的流程示意图;
图2为本发明实施例提供一种特征点对的直线交点示意图;
图3为本发明实施例提供一种判断区域示意图;
图4为现有技术提供一种特征点匹配结果判断示意图;
图5为本发明一实施例提供的特征点匹配结果处理装置结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚,下面将结合本申请具体实施例及相应的附图对本申请技术方案进行清楚、完整地描述。显然,所描述的实施例仅是本申请一部分实施例,而不是全部的实施例。基于本申请中 的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
图1为本发明一实施例提供的特征点匹配结果处理方法的流程示意图,如图1所示,
101、通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对;
首先,将第一图像和第二图像放置于同一画面中,然后将第一图像和第二图像中其中一个图像(第一图像或者第二图像)旋转180度。具体匹配时包括:
分别确定第一图像和第二图像的特征点支撑区域;
将特征点支撑区域划分为多个固定形状的子区域;
将每个子区域内包含的图像纹理信息表示为向量并组成匹配描述子;
根据匹配描述子匹配得到第一图像和第二图像中一一对应的匹配的特征点对。
由于特征点匹配算法是图像的局部特征,对平移、旋转、尺度缩放、亮度变化、遮挡和噪声等具有良好的不变性,对视觉变化、仿射变换也保持一定程度的稳定性,适用于在海量特征数据库中进行快速、准确的匹配;即使少数的几个物体也可以产生大量SIFT特征向量,速度相对较快。因此在本发明实施例中,以对第一图像和第二图像运行特征点匹配算法,其中,特征点匹配算法例如包括SIFT、SURF、GLOH等匹配描述子方法,主要是利用特征局部纹理信息进行特征点匹配,首先为特征点确定支撑区域,然后将支撑区域划分为一系列固定形状的子区域,最后将每个子区域内包含的图像纹理信息表示为向量并组成匹配描述子,根据匹配描述子匹配得到第一图像和第二图像中一一对应的匹配结果,即特征点对。
102、在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点;
由于直线方程可以表示平面中的任意一条直线,可以快速准确的得出x 与y之间存在的关系,运算速度比较快,因此本发明实施例中,对每一特征点对之间的直线计算直线方程,确定每条直线与其他直线之间的交点;
图2为本发明实施例提供一种特征点对的直线交点示意图,如图2所示,假设第一图像中的特征点(X1,Y1)与第二图像中的特征点(X2,Y2)为一对特征点对,则特征点(X1,Y1)与特征点(X2,Y2)之间的直线方程为
Figure PCTCN2018096866-appb-000001
假设两条直线方程为y=k 1x+b 1和y=k 2x+b 2,那么交点为
Figure PCTCN2018096866-appb-000002
103、根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;
在一种可选的实现方式中,根据每个交点与其他交点之间的距离,确定有效交点组包括:
在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
在得到的N-1个交点之间的距离中若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点;
将N个交点中的全部有效交点形成有效交点组。
在一种可选的实现方式中,确定有效交点组的几何中心点,包括:
若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形成的有效交点组的几何中心点;
若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
需要说明的是,在上述步骤中计算每个交点与其它所有交点的距离时,需要预先设定一个距离阈值a、第一比例阈值b%(通常可以设置为90%或80%)和第二比例阈值c%(通常可以设置为70%或60%)。
其中,距离阈值a的设置与本发明需要实现的特征点匹配结果的精度要求成反比,假设本发明需要实现的特征点匹配结果的精度要求越高,距离阈值a的值可以越小,反之,本发明需要实现的特征点匹配结果的精度要求越低,距离阈值a的值可以越大。
根据上述步骤,举例来说,随机取若干直线方程(可以是5个或10个),两两计算直线方程之间的交点,记录所有交点坐标,预先设定一个交点之间的距离阈值a,计算每个交点与其它所有交点的距离,若大于第一比例阈值b%的距离小于距离阈值a,则记录该交点为有效交点;若本次选择的直线方程产生的有效交点的数量占比(有效交点数/本次选择的交点数)小于第二比例阈值c%,则重新选择一组直线方程进行有效交点的确定;若本次选择的直线方程产生的有效交点的数量占比大于等于第二比例阈值c%,则记录所有的有效交点生成一个有效交点组,并计算该有效交点组的几何中心点。
需要说明的是,在计算有效交点组的几何中心点时,可以将各个有效交点坐标的平均值确定为几何中心点的坐标。例如,三个有效交点的坐标分别为(9,11),(10,10)和(11,9),则该三个有效交点的几何中心点的坐标为((9+10+11)/3,(11+10+9)/3)。
104、根据几何中心点确定特征点对的判断区域;
具体地,以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度。其中,长度、宽度、半径和/或对角线长度等参数可以根据本发明需要实现的特征点匹配结果的精度要求自行确定,假设本发明需要实现的特征点匹配结果的精度要求越高,长度、宽度、半径和/或对角线长度等参数可以越小,反之,本发明需要实现的特征点匹配结果的精度要求越低,长度、宽度、半径和/或对角线长度等参数可以越大。具体参考图3所示的矩形判断区域,为了实现高精度的特征点匹配结果,图3所示的矩形判断区域可以设置的比较小。
105、根据特征点对之间的直线是否通过所述判断区域确定特征点对是否 是正确的匹配结果。
本发明实施例中,假设特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;假设特征点对之间的直线不通过所述判断区域,则确定该特征点对不是正确的匹配结果。
图4为现有技术提供一种特征点匹配结果判断示意图,如图4所示,在第一图像和第二图像正向放置在同一个画面中时,如果特征点匹配结果是正确的话,匹配点对所形成的直线应该是平行的,或是呈一定规律渐变的,如图3所示,但是,这种平行的或呈一定规律渐变的判断方法不容易用数学方式表达,从而造成特征点匹配结果判断困难或判断不正确的问题。
本发明实施例中,如图3所示,将第一图像和第二图像放置于同一画面中,然后将第一图像和第二图像中的其中一个图像(第一图像或者第二图像)旋转180度。在理论上,所有正确匹配的特征点对的直线就会都相交于一个交点,因此,可以很容易通过数学方式(如特征点对之间的直线方程)计算得到两两直线方程之间的交点,通过但是在实际情况中,由于特征点坐标的细微误差,所有正确匹配的特征点对的直线相交的交点可能不是一个精确的点,所以确定了一个判断区域,当特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;当特征点对之间的直线不通过所述判断区域,则确定该特征点对不是正确的匹配结果。
本发明实施例首先将第一图像和第二图像反向放置,利用特征点匹配算法可以快色确定第一图像和第二图像之间匹配的特征点对;其次利用数学表达方式,在每一特征点对之间做直线方程,可以快速确定每条直线与其他直线之间的交点;进而根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;最后根据几何中心点确定特征点对的判断区域,从而可以根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。解决了现有的特征点匹配算法在匹配特征点完成后,只能根据肉眼检测匹配特征点的正确性而存在误差较大的问题。而本发明的技术方案通过数学表达方式可以快速地判断特征点对匹配结果的正确性,本 发明可以根据有效交点组的中心点确定一个判断区域,通过判断区域的大小可以调整匹配特征点的验证正确性,因此,通过本发明的验证方法可以对匹配特征点进行百分之百正确性的验证。
图5为本发明一实施例提供的特征点匹配结果处理装置结构示意图,如图5所示,包括:
第一确定模块,用于通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,其中,第一图像和第二图像反向放置;
第二确定模块,用于在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点;
第三确定模块,用于根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;
第四确定模块,用于根据几何中心点确定特征点对的判断区域;
第五确定模块,用于根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。
可选地,第二确定模块具体用于:
假设第一图像中的特征点(X1,Y1)与第二图像中的特征点(X2,Y2)为一对特征点对,则特征点(X1,Y1)与特征点(X2,Y2)之间的直线方程为
Figure PCTCN2018096866-appb-000003
假设两条直线方程为y=k 1x+b 1和y=k 2x+b 2,那么交点为
Figure PCTCN2018096866-appb-000004
可选地,第三确定模块具体用于:
在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
在得到的N-1个交点之间的距离中,若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点,将N个交点中的全部有效交点形成有效交点组;
若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点, 且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形成的有效交点组的几何中心点;
若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
可选地,第四确定模块具体用于:
以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度。
可选地,第五确定模块具体用于:
若特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;
若特征点对之间的直线不通过所述判断区域,则确定该特征点对是错误的匹配结果,删除错误的匹配结果。
本发明实施例所述的装置可以上述实施例所示的方法,其实现原理和技术效果不再赘述。
本发明实施例中,上述特征点匹配结果处理装置的结构中包括处理器和存储器,所述存储器用于存储支持特征点匹配结果处理装置执行上述图1所示实施例中特征点匹配结果处理方法的程序,所述处理器被配置为用于执行所述存储器中存储的程序。
所述程序包括一条或多条计算机指令,其中,所述一条或多条计算机指令供所述处理器调用执行。
本发明实施例中还提供一种计算机存储介质,其特征在于,用于储存如图1所示实施例中的特征点匹配结果处理方法所用的计算机软件指令,所述计算机软件指令包含了用于执行图1所示实施例中的特征点匹配结果处理方法所涉及的程序。
本发明实施例还提供了一种计算机存储介质,用于储存特征点匹配结果 处理装置所用的计算机软件指令,所述计算机软件指令包含了用于执行上述特征点匹配结果处理方法为特征点匹配结果处理装置所涉及的程序。
本领域内的技术人员应明白,本发明的实施例可提供为方法、系统、或计算机程序产品。因此,本发明可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中,计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器,随机存取存储器 (RAM)和/或非易失性内存等形式,如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括,但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带,磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质,可用于存储可以被计算设备访问的信息。按照本文中的界定,计算机可读介质不包括暂存电脑可读媒体(transitory media),如调制的数据信号和载波。
还需要说明的是,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括所述要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白,本申请的实施例可提供为方法、系统或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
以上所述仅为本申请的实施例而已,并不用于限制本申请。对于本领域技术人员来说,本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等,均应包含在本申请的权利要求范围之内。

Claims (10)

  1. 一种特征点匹配结果处理方法,其特征在于,包括:
    通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,其中,第一图像和第二图像反向放置;
    在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点;
    根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;
    根据几何中心点确定特征点对的判断区域;
    根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。
  2. 根据权利要求1所述的方法,其特征在于,在每一特征点对之间做直线方程,确定每条直线与其他直线之间的交点,包括:
    假设第一图像中的特征点(X1,Y1)与第二图像中的特征点(X2,Y2)为一对特征点对,则特征点(X1,Y1)与特征点(X2,Y2)之间的直线方程为
    Figure PCTCN2018096866-appb-100001
    假设两条直线方程为y=k 1x+b 1和y=k 2x+b 2,那么交点为
    Figure PCTCN2018096866-appb-100002
  3. 根据权利要求1所述的方法,其特征在于,根据每个交点与其他交点之间的距离,确定有效交点组,包括:
    在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
    在得到的N-1个交点之间的距离中若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点;
    将N个交点中的全部有效交点形成有效交点组。
  4. 根据权利要求3所述的方法,其特征在于,确定有效交点组的几何中心点,包括:
    若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点, 且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形成的有效交点组的几何中心点;
    若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
  5. 根据权利要求4所述的方法,其特征在于,根据几何中心点确定特征点对的判断区域,包括:
    以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度。
  6. 根据权利要求5所述的方法,其特征在于,根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果,包括:
    若特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;
    若特征点对之间的直线不通过所述判断区域,则确定该特征点对是错误的匹配结果,删除错误的匹配结果。
  7. 根据权利要求1所述的方法,其特征在于,通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,包括:
    分别确定第一图像和第二图像的特征点支撑区域;
    将特征点支撑区域划分为多个固定形状的子区域;
    将每个子区域内包含的图像纹理信息表示为向量并组成匹配描述子;
    根据匹配描述子匹配得到第一图像和第二图像中一一对应的匹配的特征点对。
  8. 一种特征点匹配结果处理装置,其特征在于,包括:
    第一确定模块,用于通过特征点匹配算法确定第一图像和第二图像之间匹配的特征点对,其中,第一图像和第二图像反向放置;
    第二确定模块,用于在每一特征点对之间做直线方程,确定每条直线与 其他直线之间的交点;
    第三确定模块,用于根据每个交点与其他交点之间的距离,确定有效交点组以及有效交点组的几何中心点;
    第四确定模块,用于根据几何中心点确定特征点对的判断区域;
    第五确定模块,用于根据特征点对之间的直线是否通过所述判断区域确定特征点对是否是正确的匹配结果。
  9. 根据权利要求8所述的装置,其特征在于,第三确定模块具体用于:
    在随机选择的N个交点中,根据每个交点的坐标,计算每个交点与其他N-1个交点之间的距离;
    在得到的N-1个交点之间的距离中,若存在M个距离小于预设的距离阈值,且M/N-1大于等于第一比例阈值,则确定该交点为有效交点,将N个交点中的全部有效交点形成有效交点组;
    若N个交点中的全部有效交点形成的有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时,根据K个有效交点确定该K个有效交点形成的有效交点组的几何中心点;
    若K/N小于第二比例阈值,则重新随机选择的N个交点中确定有效交点组,直至重新确定该有效交点组中存在K个有效交点,且K/N大于等于第二比例阈值时为止。
  10. 根据权利要求9所述的装置,其特征在于,第四确定模块具体用于:
    以几何中心点为中心,根据预设的参数设置特征点对的判断区域,所述判断区域包括矩形区域、圆形区域或椭圆区域,对应的参数包括长度、宽度、半径和/或对角线长度;
    第五确定模块具体用于:
    若特征点对之间的直线通过所述判断区域,则确定该特征点对是正确的匹配结果;
    若特征点对之间的直线不通过所述判断区域,则确定该特征点对是错误的匹配结果,删除错误的匹配结果。
PCT/CN2018/096866 2017-08-01 2018-07-24 特征点匹配结果处理方法和装置 Ceased WO2019024723A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/620,790 US11100361B1 (en) 2017-08-01 2018-07-24 Method and apparatus for processing feature point matching result

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710648107.0A CN107480710B (zh) 2017-08-01 2017-08-01 特征点匹配结果处理方法和装置
CN201710648107.0 2017-08-01

Publications (1)

Publication Number Publication Date
WO2019024723A1 true WO2019024723A1 (zh) 2019-02-07

Family

ID=60598477

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/096866 Ceased WO2019024723A1 (zh) 2017-08-01 2018-07-24 特征点匹配结果处理方法和装置

Country Status (3)

Country Link
US (1) US11100361B1 (zh)
CN (1) CN107480710B (zh)
WO (1) WO2019024723A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107480710B (zh) 2017-08-01 2020-05-22 歌尔股份有限公司 特征点匹配结果处理方法和装置
CN112183596B (zh) * 2020-09-21 2022-10-11 湖北大学 结合局部网格约束和几何约束的直线段匹配方法与系统
CN116245858A (zh) * 2023-03-15 2023-06-09 西安元智系统技术有限责任公司 一种文物劣化识别方法及装置
CN116342907B (zh) * 2023-03-23 2024-09-06 苏州大学 交点增强的图像间直线段匹配方法和系统
CN119091164A (zh) * 2023-06-06 2024-12-06 北京字跳网络技术有限公司 图像处理方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102385750A (zh) * 2011-06-22 2012-03-21 清华大学 基于几何关系的直线匹配方法及系统
JP2013178656A (ja) * 2012-02-28 2013-09-09 Topcon Corp 画像処理装置、画像処理方法および画像処理用のプログラム
CN104778682A (zh) * 2014-01-14 2015-07-15 三星泰科威株式会社 对特征点采样的方法、图像匹配方法和图像匹配设备
CN107480710A (zh) * 2017-08-01 2017-12-15 歌尔股份有限公司 特征点匹配结果处理方法和装置

Family Cites Families (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS60221882A (ja) * 1984-04-18 1985-11-06 Nec Corp 本人照合方式
JPS60221880A (ja) * 1984-04-18 1985-11-06 Nec Corp 本人照合方式
JP4167390B2 (ja) * 2000-11-20 2008-10-15 日本電気株式会社 物体照合方法,物体照合装置,およびそのプログラムを記録した記録媒体
US7373275B2 (en) * 2003-03-14 2008-05-13 Rudolph Technologies, Inc. System and method of point matching measured positions to template positions
WO2006034256A2 (en) * 2004-09-17 2006-03-30 Cyberextruder.Com, Inc. System, method, and apparatus for generating a three-dimensional representation from one or more two-dimensional images
US7412427B2 (en) * 2006-01-27 2008-08-12 Microsoft Corporation Object instance recognition using feature symbol triplets
US7991189B2 (en) * 2007-03-28 2011-08-02 Sharp Kabushiki Kaisha Image processing apparatus, image forming apparatus, image processing system, and image processing method
US8103108B2 (en) * 2007-05-01 2012-01-24 Sharp Kabushiki Kaisha Image processing apparatus, image forming apparatus, image processing system, and image processing method
CN201378656Y (zh) 2009-03-06 2010-01-06 北京华旗资讯数码科技有限公司 一种具有两种标准接触件的存储卡
CN101726296B (zh) * 2009-12-22 2013-10-09 哈尔滨工业大学 空间机器人视觉测量、路径规划、gnc一体化仿真系统
US8532402B2 (en) * 2011-06-22 2013-09-10 The Boeing Company Image registration
CN102708368B (zh) * 2012-05-04 2014-01-15 湖南大学 一种基于机器视觉的生产线上瓶体的定位方法
CN103020938B (zh) * 2012-12-14 2015-10-07 北京经纬恒润科技有限公司 一种基于加权平均法的空域图像缝合方法和系统
CN103150737A (zh) * 2013-01-18 2013-06-12 西北工业大学 一种适用于空间绳系机器人的实时空间目标特征点跟踪方法
US8929602B2 (en) * 2013-01-31 2015-01-06 Seiko Epson Corporation Component based correspondence matching for reconstructing cables
CN104036542B (zh) * 2014-05-21 2017-01-25 北京信息科技大学 一种基于空间光线聚集性的像面特征点匹配方法
US9634486B2 (en) * 2014-07-09 2017-04-25 Qualcomm Incorporated Dynamic power rail control for clusters of loads
US10198865B2 (en) * 2014-07-10 2019-02-05 Seiko Epson Corporation HMD calibration with direct geometric modeling
CN105444696B (zh) * 2015-12-30 2018-04-24 天津大学 一种基于透视投影直线测量模型的双目匹配方法及其应用
US10109055B2 (en) * 2016-11-21 2018-10-23 Seiko Epson Corporation Multiple hypotheses segmentation-guided 3D object detection and pose estimation
CN106960179B (zh) * 2017-02-24 2018-10-23 北京交通大学 轨道交通线路环境安全智能监测方法和装置
JP6976733B2 (ja) * 2017-06-14 2021-12-08 キヤノン株式会社 画像処理装置、画像処理方法、およびプログラム

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102385750A (zh) * 2011-06-22 2012-03-21 清华大学 基于几何关系的直线匹配方法及系统
JP2013178656A (ja) * 2012-02-28 2013-09-09 Topcon Corp 画像処理装置、画像処理方法および画像処理用のプログラム
CN104778682A (zh) * 2014-01-14 2015-07-15 三星泰科威株式会社 对特征点采样的方法、图像匹配方法和图像匹配设备
CN107480710A (zh) * 2017-08-01 2017-12-15 歌尔股份有限公司 特征点匹配结果处理方法和装置

Also Published As

Publication number Publication date
CN107480710B (zh) 2020-05-22
CN107480710A (zh) 2017-12-15
US11100361B1 (en) 2021-08-24
US20210279504A1 (en) 2021-09-09

Similar Documents

Publication Publication Date Title
WO2019024723A1 (zh) 特征点匹配结果处理方法和装置
CN112085033B (zh) 一种模板匹配方法、装置、电子设备及存储介质
Zheng et al. A general and simple method for camera pose and focal length determination
WO2020098296A1 (zh) 一种图像检索方法及装置
US8204714B2 (en) Method and computer program product for finding statistical bounds, corresponding parameter corners, and a probability density function of a performance target for a circuit
TW202011264A (zh) 資訊的檢測方法、裝置及設備
CN114610831A (zh) 高精地图车道的实时更新方法及装置
CN114371475A (zh) 一种优化标定参数的方法、系统、设备及计算机存储介质
CN116188960A (zh) 基于关键点识别的仪表智能读数方法、装置、设备和介质
WO2017050083A1 (zh) 一种元件识别方法及装置
CN115096286A (zh) 地图合并方法、装置、机器人、存储介质和程序产品
CN103208112A (zh) 多个相机的多部分对应器
CN114998755A (zh) 在遥感图像中地标匹配的方法及装置
US12153175B2 (en) Method and terminal device for processing Positron Emission Tomography data
CN111552751A (zh) 三维地标控制点生成及应用方法、生成及应用装置
US20210166053A1 (en) Merging object detections using graphs
CN115937003A (zh) 图像处理方法、装置、终端设备和可读存储介质
CN112819861B (zh) 点云的运动补偿方法、装置与计算机可读存储介质
US8630477B2 (en) Electronic device and method for outputting measurement data
Kanaeva et al. Camera Pose and Focal Length Estimation Using Regularized Distance Constraints.
CN113536025A (zh) 兴趣点的招牌朝向确定方法、装置、电子设备和存储介质
CN118115474A (zh) 焊接质量检测轨迹点法向量估算方法、装置、设备及介质
CN110827259B (zh) 一种用于目标关键点检测的样本增广的方法与设备
Akizuki et al. DPN-LRF: A local reference frame for robustly handling density differences and partial occlusions
CN110020354A (zh) 点击热力图异常检测方法及装置

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: 18841107

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: 18841107

Country of ref document: EP

Kind code of ref document: A1