CN113139626B - Template matching method and device, electronic equipment and computer-readable storage medium - Google Patents
Template matching method and device, electronic equipment and computer-readable storage medium Download PDFInfo
- Publication number
- CN113139626B CN113139626B CN202110682462.6A CN202110682462A CN113139626B CN 113139626 B CN113139626 B CN 113139626B CN 202110682462 A CN202110682462 A CN 202110682462A CN 113139626 B CN113139626 B CN 113139626B
- Authority
- CN
- China
- Prior art keywords
- image
- feature vector
- pyramid
- determining
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation 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
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The application provides a template matching method, a template matching device, electronic equipment and a computer-readable storage medium; the method comprises the following steps: determining an edge feature vector of the template image; decomposing an image to be matched into a pyramid image sequence; traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image; and determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image. By the method and the device, the calculation complexity of template matching can be reduced, and the precision of template matching is improved.
Description
Technical Field
The present application relates to image processing technologies, and in particular, to a template matching method and apparatus, an electronic device, and a computer-readable storage medium.
Background
Template matching is the most basic and commonly used matching method in image processing technology and is one of the core algorithms for machine vision applications. Template matching is used as a first step of machine vision application, so that the precision and speed of template matching determine the application prospect of the machine vision application.
Template matching is widely applied to the fields of target positioning, target identification, image registration, robot guidance and the like. The position coordinates, angles, scales and matching scores of the targets can be returned through template matching, positioning guidance can be performed according to the position information of the template matching, and whether the matched targets are abnormal in the current scene or not can be judged according to the actual template matching result; judging whether the workpieces on the production line are missed according to the fraction of template matching, judging whether the directions of the workpieces are consistent according to the angle of template matching, and judging whether products of other specifications are mixed according to the scale of template matching.
In the related art, the template matching scheme generally includes a template matching scheme based on edge features, a template matching scheme based on correlation, and a template matching scheme based on feature points; however, the template matching scheme provided by the related art has the problems of high computational complexity, poor template matching accuracy and the like.
Disclosure of Invention
The embodiment of the application provides a template matching method, a template matching device, electronic equipment and a computer-readable storage medium, which can reduce the calculation complexity of template matching and improve the precision of template matching.
The technical scheme of the embodiment of the application is realized as follows:
the embodiment of the application provides a template matching method, which comprises the following steps:
determining an edge feature vector of the template image;
decomposing an image to be matched into a pyramid image sequence;
traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image;
and determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image.
In some embodiments, said traversing a search over each pyramid image in said sequence of pyramid images in top-down order comprises:
and sliding a window on each pyramid image by taking pixel points as sliding units according to the sequence from the topmost pyramid image to the bottommost pyramid image in the pyramid image sequence.
In some embodiments, the determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image comprises:
performing the following operations for each pyramid image:
acquiring a characteristic vector of a region image corresponding to the window in the pyramid image;
determining a vector difference of the edge feature vector and a feature vector of the region image;
acquiring a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction;
and determining the sum of the first projection length and the second projection length, and using the sum of the first projection length and the second projection length as the similarity of the edge feature vector and the feature vector of the region image.
In some embodiments, the determining, based on the similarity between the edge feature vector and the feature vector on each pyramid image, a target image on the image to be matched that matches the template image includes:
performing the following operations for each pyramid image:
determining a feature vector with the similarity smaller than a similarity threshold value with the edge feature vector as a target feature vector;
and if the number of the target feature vectors in the region image is greater than a number threshold, determining that the region image is the target image matched with the template image.
In some embodiments, the determining the edge feature vector of the template image comprises:
determining gradient information of each pixel point in the template image, and determining an edge feature vector of the determined template image based on the gradient information.
In some embodiments, the determining gradient information of each pixel point in the template image, and the determining an edge feature vector based on the gradient information includes:
inputting the template image into a gradient model, and determining gradient information of each pixel point in the template image according to the output of the gradient model; the gradient information comprises a gradient direction and a gradient amplitude;
determining pixel points of which the gradient amplitudes are larger than the amplitude threshold value as edge pixel points;
and determining the feature vector of the edge pixel point as the edge feature vector.
In some embodiments, the decomposing the image to be matched into a sequence of pyramid images includes:
sampling the image to be matched in a gradient down-sampling mode to obtain at least two pyramid images with different resolutions corresponding to the image to be matched; the at least two pyramid images form the pyramid image sequence.
In some embodiments, before determining the similarity between the edge feature vector and the feature vector on each pyramid image, the method further comprises:
and respectively carrying out normalization processing on the edge feature vector and the feature vector on each pyramid image, so that the edge feature vector and the feature vector on each pyramid image can be represented on a circle with the radius of 1.
The embodiment of the application provides a template matching device, includes:
the edge feature vector determining module is used for determining an edge feature vector of the template image;
the decomposition module is used for decomposing the image to be matched into a pyramid image sequence;
a similarity determining module, configured to perform traversal search on each pyramid image in the pyramid image sequence according to an order from top to bottom, and determine, based on a projection length of a difference between the edge feature vector and a feature vector on each pyramid image, a similarity between the edge feature vector and the feature vector on each pyramid image;
and the matching module is used for determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image.
In some embodiments, the similarity determining module is configured to slide a window on each pyramid image by using a pixel point as a sliding unit according to an order from a top-most pyramid image to a bottom-most pyramid image in the pyramid image sequence.
In some embodiments, the similarity determining module is configured to perform the following operations for each pyramid image:
acquiring a characteristic vector of a region image corresponding to the window in the pyramid image;
determining a vector difference of the edge feature vector and a feature vector of the region image;
acquiring a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction;
and determining the sum of the first projection length and the second projection length, and using the sum of the first projection length and the second projection length as the similarity of the edge feature vector and the feature vector of the region image.
In some embodiments, the similarity determining module is configured to perform the following operations for each pyramid image:
determining a feature vector with the similarity smaller than a similarity threshold value with the edge feature vector as a target feature vector;
and if the number of the target feature vectors in the region image is greater than a number threshold, determining that the region image is the target image matched with the template image.
In some embodiments, the edge feature vector determining module is configured to determine gradient information of each pixel point in the template image, and determine the edge feature vector of the determined template image based on the gradient information.
In some embodiments, the edge feature vector determining module is configured to input the template image to a gradient model, and determine gradient information of each pixel point in the template image according to an output of the gradient model; the gradient information comprises a gradient direction and a gradient amplitude;
determining pixel points of which the gradient amplitudes are larger than the amplitude threshold value as edge pixel points;
and determining the feature vector of the edge pixel point as an edge feature vector.
In some embodiments, the decomposition module is configured to sample the image to be matched in a manner of gradient down-sampling to obtain at least two pyramid images with different resolutions corresponding to the image to be matched; the at least two pyramid images form the pyramid image sequence.
In some embodiments, the similarity determination module is further configured to perform normalization processing on the edge feature vector and the feature vector on each pyramid image, respectively, so that the edge feature vector and the feature vector on each pyramid image can be represented on a circle with a radius of 1.
An embodiment of the present application provides an electronic device, including:
a memory for storing executable instructions;
and the processor is used for realizing the template matching method provided by the embodiment of the application when the executable instructions stored in the memory are executed.
The embodiment of the present application provides a computer-readable storage medium, which stores executable instructions and is used for implementing the template matching method provided by the embodiment of the present application when being executed by a processor.
The template matching method provided by the embodiment of the application determines the edge feature vector of the template image; decomposing an image to be matched into a pyramid image sequence; traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image; and determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image. Therefore, the template matching method provided by the embodiment of the application can determine the similarity between the edge feature vector of the template image and the feature vector on the pyramid image corresponding to the image to be matched based on the projection length of the difference between the edge feature vector of the template image and the feature vector on the pyramid image corresponding to the image to be matched, the expression form of the similarity is simple, only addition operation and subtraction operation are involved in calculating the similarity, and the calculation complexity is low. In addition, the template matching method provided by the embodiment of the application can realize template matching only based on the projection length of the difference between the edge feature vector of the template image and the feature vector on the pyramid image corresponding to the image to be matched, does not depend on the parameters such as the angle step length of search, and improves the precision of template matching.
Drawings
Fig. 1 is a schematic architecture diagram of a template matching system provided in an embodiment of the present application;
fig. 2 is a schematic architecture diagram of a terminal device provided in an embodiment of the present application;
fig. 3 is a schematic flowchart of a template matching method provided in an embodiment of the present application;
FIG. 4 is a schematic diagram of a template image provided by an embodiment of the present application;
FIG. 5 is a schematic diagram of an edge image provided by an embodiment of the present application;
FIG. 6A is a schematic diagram of a pyramid image sequence provided by an embodiment of the present application;
FIG. 6B is a schematic diagram of a feature pyramid of an image provided in an embodiment of the present application;
FIG. 7 is a schematic view of a sliding window provided by an embodiment of the present application;
FIG. 8 is a schematic diagram of an alternative process flow for determining similarity between the edge feature vector and the feature vector on the pyramid image according to the embodiment of the present application;
FIG. 9 is a diagram of normalized edge feature vectors and feature vectors of a region image according to an embodiment of the present disclosure;
FIG. 10 is a view showing a view for use in the related art
A schematic diagram representing the similarity of two feature vectors;
fig. 11 is a schematic diagram illustrating similarity between two feature vectors according to a projection length of a vector difference between the two feature vectors on a coordinate axis according to an embodiment of the present application.
Detailed Description
In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.
In the following description, references to the terms "first \ second \ third" are only to distinguish similar objects and do not denote a particular order, but rather the terms "first \ second \ third" are used to interchange specific orders or sequences, where appropriate, so as to enable the embodiments of the application described herein to be practiced in other than the order shown or described herein. In the following description, the term "plurality" referred to means at least two.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Before further detailed description of the embodiments of the present application, terms and expressions referred to in the embodiments of the present application will be described, and the terms and expressions referred to in the embodiments of the present application will be used for the following explanation.
1) The template image is a known image.
2) Template matching, namely searching a target in a large image, knowing that the large image has the target to be matched and the target has the same size, direction and image elements as the template image, finding the target in the large image through a certain algorithm and determining the coordinate position of the target.
3) The image pyramid is a kind of multi-scale expression of an image, and is a structure for interpreting an image with multiple resolutions. A pyramid corresponding to an image is a set of images arranged in a pyramid shape with progressively lower resolutions and derived from the same original image. Comparing the images of one layer with the images of one layer into a pyramid, wherein the higher the level is, the smaller the image is, and the lower the resolution is; conversely, the lower the hierarchy, the larger the image and the higher the resolution.
4) The edge is the most basic characteristic of the image and refers to a set of pixels with step change or roof-shaped change in the gray level of the pixels in the image; the edges are associated with discontinuities in the image brightness or first derivative of the image brightness and thus appear as step edges and line edges.
5) Database (Database): similar to an electronic file cabinet, namely a place for storing electronic files, a user can perform operations of adding, inquiring, updating, deleting and the like on data in the files. A database is also to be understood as a collection of data that are stored together in a manner that can be shared with a plurality of users, with as little redundancy as possible, independent of the application. In embodiments of the present application, the database may store data for model training.
The embodiment of the application provides a template matching method, a template matching device, electronic equipment and a computer-readable storage medium, which can reduce the calculation complexity of template matching and improve the precision of template matching. An exemplary application of the electronic device provided in the embodiment of the present application is described below, and the electronic device provided in the embodiment of the present application may be implemented as various types of terminal devices, and may also be implemented as a server.
Referring to fig. 1, fig. 1 is an architecture diagram of a template matching system 100 provided in an embodiment of the present application, a terminal device 400 is connected to a server 200 through a network 300, and the server 200 is connected to a database 500, where the network 300 may be a wide area network or a local area network, or a combination of the two.
In some embodiments, taking an example that the electronic device implementing the template matching method is a terminal device, the template matching method provided in the embodiments of the present application may be implemented by the terminal device. For example, the terminal device 400 runs the client 410, and the client 410 may be a client for performing template matching.
When template matching is needed, the client 410 acquires a template image and then determines an edge feature vector of the template image; the client 410 decomposes the image to be matched into a pyramid image sequence; traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image; and determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image.
In some embodiments, taking an example that the electronic device implementing the template matching method is a server, the template matching method provided in the embodiments of the present application may be cooperatively implemented by the server and the terminal device. For example, the server 200 obtains a sample image from the database 500, and labels gradient information of each pixel point in the sample image. And then, training a gradient model according to the sample image and the gradient information of each pixel point in the sample image.
The server 200 acquires the template image from the client 410. Then, the server 200 determines an edge feature vector of the template image; decomposing an image to be matched into a pyramid image sequence; traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image; determining a target image matched with the template image on the image to be matched based on the similarity of the edge feature vector and the feature vector on each pyramid image; the server 200 sends the target image to the client 410.
In some embodiments, the terminal device 400 or the server 200 may implement the template matching method provided by the embodiments of the present application by running a computer program, for example, the computer program may be a native program or a software module in an operating system; can be a local (Native) Application program (APP), i.e. a program that needs to be installed in an operating system to run; or may be an applet, i.e. a program that can be run only by downloading it to the browser environment; but also an applet that can be embedded into any APP. In general, the computer programs described above may be any form of application, module or plug-in.
In some embodiments, the server 200 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a Cloud server providing basic Cloud computing services such as a Cloud service, a Cloud database, Cloud computing, a Cloud function, Cloud storage, a web service, Cloud communication, a middleware service, a domain name service, a security service, a CDN, and a big data and artificial intelligence platform, where Cloud Technology (Cloud Technology) refers to a hosting Technology for unifying resources of hardware, software, a network, and the like in a wide area network or a local area network to implement computing, storage, processing, and sharing of data. The terminal device 400 may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, and the like. The terminal device and the server may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present application is not limited.
Taking the electronic device provided in the embodiment of the present application as an example for illustration, it can be understood that, for the case where the electronic device is a server, parts (such as the user interface, the presentation module, and the input processing module) in the structure shown in fig. 2 may be default. Referring to fig. 2, fig. 2 is a schematic structural diagram of a terminal device 400 provided in an embodiment of the present application, where the terminal device 400 shown in fig. 2 includes: at least one processor 410, memory 450, at least one network interface 420, and a user interface 430. The various components in the terminal 400 are coupled together by a bus system 440. It is understood that the bus system 440 is used to enable communications among the components. The bus system 440 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 440 in fig. 2.
The Processor 410 may be an integrated circuit chip having Signal processing capabilities, such as a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like, wherein the general purpose Processor may be a microprocessor or any conventional Processor, or the like.
The user interface 430 includes one or more output devices 431, including one or more speakers and/or one or more visual displays, that enable the presentation of media content. The user interface 430 also includes one or more input devices 432, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.
The memory 450 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, and the like. Memory 450 optionally includes one or more storage devices physically located remote from processor 410.
The memory 450 includes either volatile memory or nonvolatile memory, and may include both volatile and nonvolatile memory. The nonvolatile Memory may be a Read Only Memory (ROM), and the volatile Memory may be a Random Access Memory (RAM). The memory 450 described in embodiments herein is intended to comprise any suitable type of memory.
In some embodiments, memory 450 is capable of storing data, examples of which include programs, modules, and data structures, or a subset or superset thereof, to support various operations, as exemplified below.
An operating system 451, including system programs for handling various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and handling hardware-based tasks;
a network communication module 452 for communicating to other computing devices via one or more (wired or wireless) network interfaces 420, exemplary network interfaces 420 including: bluetooth, wireless compatibility authentication (WiFi), and Universal Serial Bus (USB), etc.;
a presentation module 453 for enabling presentation of information (e.g., user interfaces for operating peripherals and displaying content and information) via one or more output devices 431 (e.g., display screens, speakers, etc.) associated with user interface 430;
an input processing module 454 for detecting one or more user inputs or interactions from one of the one or more input devices 432 and translating the detected inputs or interactions.
In some embodiments, the apparatus provided in the embodiments of the present application may be implemented in software, and fig. 2 illustrates a template matching apparatus 455 stored in the memory 450, which may be software in the form of programs and plug-ins, and may include the following software modules: an edge feature vector determination module 4551, a decomposition module 4552, a similarity determination module 4553 and a matching module 4554, which are logical and thus may be arbitrarily combined or further split depending on the functions implemented. The functions of the respective modules will be explained below.
In the related art, a template matching scheme based on feature points is to use the feature points as matched feature information, feature points need to be found in a template image and an image to be matched, the feature points need to meet the characteristics of corner points, and feature description needs to be performed on the feature points for subsequent template matching. However, for some scenes in which the template images only have linear and circular features, the template matching method based on the feature points cannot extract the feature points, so that subsequent template matching operation cannot be performed. Moreover, the feature point information of the template image and the image to be matched needs to be calculated, the perspective transformation relation between the template image and the image to be matched is established, and the calculation complexity is high; if real-time matching or fast matching needs to be met, the hardware performance of the electronic equipment is required to be higher.
In the correlation technique, the template matching scheme based on the correlation has huge calculation amount on the correlation, can not achieve higher matching precision, and can not be applied to the application scene of local illumination transformation; therefore, the correlation-based template matching scheme is not suitable for a scenario with high requirements on template matching performance.
In the related art, the generalized Hough transform-based method realizes the matching of edge features, and when template matching is further performed, the generalized Hough transform calculates matching information in an accumulated mode based on candidate targets, so that the method depends on the angle step length of search, and the precision of template matching cannot be ensured.
In other schemes of template matching based on edge features, when the similarity of the edge features is calculated, the included angle fraction of the edge feature vector is calculated by utilizing the cosine value of the angle; when the cosine value of the angle is calculated, penalty operation and division operation are included, and the calculation complexity is high.
The embodiment of the application provides a template matching method, which can at least solve the problem.
The template matching method provided by the embodiment of the present application will be described below with reference to exemplary applications and implementations of the electronic device provided by the embodiment of the present application.
Referring to fig. 3, fig. 3 is a schematic flowchart of an alternative template matching method provided in an embodiment of the present application, which will be described with reference to the steps shown in fig. 3.
Step S101, determining an edge feature vector of the template image.
In some embodiments, gradient information of each pixel point in the template image is determined, and an edge feature vector of the determined template image is determined based on the gradient information.
In some embodiments, the gradient of the image indicates the speed of change of the image, reflecting edge information of the image. Therefore, for the edge part of the image, the gray value change is large, and the gradient value change is also large; for smoother parts of the image, the gray value variation is smaller and the gradient value variation is also smaller. Based on this, the embodiment of the application detects the discontinuity of the template image by determining the gradient information of each pixel point in the template image, and further determines the edge feature vector of the template image.
In some embodiments, the template image may be input to a pre-trained gradient model, and gradient information of each pixel point in the template image is determined according to an output of the gradient model; the gradient information comprises a gradient direction and a gradient amplitude; determining pixel points of which the gradient amplitudes are larger than the amplitude threshold value as edge pixel points; and determining the feature vector of the edge pixel point as an edge feature vector. The gradient threshold value can be flexibly set according to the actual application scene. The gradient information of a pixel point can be expressed as (m)x，my) The gradient information may also be normalized to obtain normalized gradient information a =
。
Taking the template image shown in fig. 4 as an example, the edge image shown in fig. 5 can be obtained by performing edge detection on the template image shown in fig. 4; in specific implementation, the edge detection may be performed on the template image based on an existing algorithm or technology, and the embodiment of the present application is not limited. The edge feature vector of the template image is a feature vector corresponding to the edge image in the template image.
In some embodiments, the process of determining the edge feature vector of the template image in step S101 may also be referred to as template training.
And step S102, decomposing the image to be matched into a pyramid image sequence.
In some embodiments, the image to be matched is sampled in a gradient down-sampling manner to obtain at least two pyramid images with different resolutions corresponding to the image to be matched; the at least two pyramid images form the pyramid image sequence. As an example, the resolution of each pyramid image in the sequence of pyramid images is different.
In some embodiments, a schematic diagram of a sequence of pyramid images, as shown in fig. 6A, the bottom layer of the sequence of pyramid images is a high resolution representation of the images to be matched and the top layer of the sequence of pyramid images is a low resolution representation of the images to be matched. When moving to the upper layer of the pyramid of the image sequence, the size and resolution of the image are reduced, wherein the size of the upper layer image of the pyramid image sequence is 1/4 of the size of the previous layer image, the number of pyramid images included in the pyramid image sequence can be 0,1,2 … … N, and the number of layers of the corresponding pyramid is 0,1,2 … … N; fig. 6A takes the number of pyramid layers as 4 as an example, and decomposes an image to be matched into an image to be matched 1, an image to be matched 2, an image to be matched 3, and an image to be matched 4; the resolution of the image 1 to be matched is the lowest, and the resolution of the image 4 to be matched is the highest. As an example, the feature pyramid diagram of the image, as shown in fig. 6B, the image resolution presented by the bottom layer of the feature pyramid is the highest, and the image resolution presented by the top layer of the feature pyramid is the lowest.
Step S103, performing traversal search on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image.
In some embodiments, traversing the search over each pyramid image in the sequence of pyramid images in top-down order may be: and sliding a window on each pyramid image by taking pixel points as sliding units according to the sequence from the topmost pyramid image to the bottommost pyramid image in the pyramid image sequence. On the bottom pyramid image of the pyramid image sequence, a more accurate pose corresponding to the template image can be matched.
Taking the pyramid image sequence shown in fig. 6A as an example, firstly, the window is slid on the image to be matched 1, the image to be matched 2, the image to be matched 3 and the image to be matched 4 in sequence by taking the pixel point as a unit, and the size of the window is the same as that of the template image. As shown in fig. 7, the schematic diagram of the sliding window starts to slide the window from the starting point on the image to be matched, and the offset of two adjacent sliding windows is 1 pixel point. Through sliding creation, the similarity between the area image on the image to be matched corresponding to the window and the template image can be calculated.
In some embodiments, an optional process flow for determining similarity between the edge feature vector and the feature vector on the pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on the pyramid image for each pyramid image may be as shown in fig. 8, including:
step S103a, obtaining a feature vector of the area image corresponding to the window in the pyramid image.
In some embodiments, the number of feature vectors of the region image corresponding to the window may be the same as the number of edge feature vectors of the template image in step S101.
Step S103b, a vector difference between the edge feature vector and the feature vector of the region image is determined.
In some embodiments, the normalization process may be performed on the edge feature vector and the feature vector of the region image, respectively, first, so that both the edge feature vector and the feature vector of the region image can be represented on a circle with a radius of 1. As shown in fig. 9, the normalized edge feature vector is represented by (x 1, y 1), and the normalized region image feature vector is represented by (x 2, y 2), so that the vector difference between the edge feature vector and the region image feature vector can be represented by (x 2-x 1, y 2-y 1).
Step S103c, a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction are obtained.
In some embodiments, the first direction may be an X-axis direction in a coordinate system, and the second direction may be a Y-axis direction in the coordinate system.
In some embodiments, if the vector difference of the edge feature vector and the feature vector of the region image is represented as (X2-X1, y 2-y 1), the first projection length of the vector difference in the X-axis direction is (X2-X1, y 2-y 1)
A second projection length of the vector difference in the Y-axis direction is
。
Step S103d, determining a sum of the first projection length and the second projection length, and using the sum of the first projection length and the second projection length as a similarity between the edge feature vector and the feature vector of the region image.
In some embodiments, if the first projection length of the vector difference in the X-axis direction is
A second projection length of the vector difference in the Y-axis direction is
The similarity between the edge feature vector and the feature vector of the region image is equal to
+
。
In the embodiment of the application, the similarity between the edge feature vector and the feature vector of the region image is represented by the sum of the projection lengths of the vector difference between the edge feature vector and the feature vector of the region image in the X axis and the Y axis. When calculating the projection lengths of the vector difference between the edge feature vector and the feature vector of the region image in the X axis and the Y axis, only addition operation and subtraction operation are involved, the calculation amount is small, and the calculation complexity is low.
And step S104, determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image.
In some embodiments, the following operations are performed separately for each pyramid image: determining the feature vector of the area image with the similarity smaller than the similarity threshold value with the edge feature vector on the pyramid image as a target feature vector; that is, the smaller the calculated similarity value is, the higher the similarity characterizing the two feature vectors is. And if the number of the target feature vectors in the region image is greater than a number threshold, determining that the region image is the target image matched with the template image.
In specific implementation, each pyramid image comprises a plurality of region images, and the similarity between the edge feature vector of the template image and the feature vector of each region image is calculated; and when the similarity is greater than a preset similarity threshold, determining the feature vector of the regional image as a target feature vector. The similarity threshold may be flexibly set according to an actual application scenario and an application requirement, and for example, the similarity threshold may be set to 0.8. And if the number of the target feature vectors in one region image is greater than a number threshold, determining that the region image is the target image matched with the template image. As an example, if the number of edge feature vectors of a template image is N, the number threshold may be set to 0.8N, and if the number of target feature vectors in a region image is greater than 0.8N, the region image is determined to be a target image matching the template image. The quantity threshold value can be flexibly set according to the actual application scene and the application requirement.
In some embodiments, steps S102 to S104 may also become a process of template matching.
The template matching method provided by the embodiment of the application determines the edge feature vector of the template image; decomposing an image to be matched into a pyramid image sequence; traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and determining the similarity between the edge feature vector and the feature vector on each pyramid image based on the projection length of the difference between the edge feature vector and the feature vector on each pyramid image; and determining a target image matched with the template image on the image to be matched based on the similarity between the edge feature vector and the feature vector on each pyramid image. According to the template matching method provided by the embodiment of the application, the template matching scheme based on the shape or the outline is adopted, so that the template matching algorithm can achieve very high matching precision on the rotation and scale change of the template image, the angle precision can reach 0.02 degree, and the pixel precision can reach 1/40 pixels. Determining the similarity between the edge feature vector and the feature vector on each pyramid image according to the projection length of the difference between the edge feature vector and the feature vector on each pyramid image, and determining a target image matched with the template image on the image to be matched based on the similarity; the similarity is determined by using the projection length, so that the expression form of the similarity is simpler. When the projection length is calculated, only addition operation and subtraction operation are involved, so that the calculation amount of template matching is small, and the calculation complexity is low; the method has low hardware requirement on the electronic equipment for realizing the template matching method, and can be suitable for scenes for performing template matching in real time or quickly.
In the related art, the similarity measure needs to reflect the change of the two feature vectors in the distance or the included angle, and the smaller the change, the closer the two feature vectors are, the higher the similarity is (or the higher the similarity score is). Taking an angle as an example, assuming that two eigenvectors are respectively represented by (x 1, y 1) and (x 2, y 2), the included angle between the two eigenvectors is
The similarity of two feature vectors can be used
Wherein, in the step (A),
the similarity between the template image and the image to be matched can be expressed as:
Similarity=
=
=
Wherein, the similarity between two feature points can be determined by the following formula:
similarity= x1* x2+ y1* y2。
in the related art, when the similarity is calculated by the euclidean distance, the similarity can be determined by the following formula:
similarity=
。
in the related art, when the similarity of two eigenvectors is determined, multiplication operation is involved, and even root operation is involved; in the embodiment of the present application, only addition and subtraction are involved in determining the similarity of two feature vectors. The calculation amount of multiplication and root operation is obviously higher than that of addition and subtraction in the embodiment of the application, correspondingly, the calculation speed of multiplication and root operation is obviously lower than that of addition and subtraction in the embodiment of the application, and the requirement of multiplication and root operation on the hardware performance of the electronic equipment is also higher than that of addition and subtraction on the hardware performance of the electronic equipment.
In order to verify the accuracy of the template matching method provided by the embodiment of the application, the included angle between the eigenvector (1, 0) and (x 2, y 2) is used
Based on the use in the related art
A schematic diagram representing the similarity of two feature vectors is shown in FIG. 10, with the abscissa representing
The ordinate represents the value of the similarity; it can be seen that
When the temperature is higher than 0 degree,
the similarity characterizing the two feature vectors is the highest. In the embodiment of the present application, the projection length of the vector difference between the two feature vectors on the coordinate axis represents the similarity between the two feature vectors, as shown in fig. 11, the abscissa represents the angle between the two feature vectors, and the ordinate represents the value of the similarity, as can be seen from the above description, in the embodiment of the present application, the projection length of the vector difference between the two feature vectors on the coordinate axis represents the similarity between the two feature vectors
0 degree, vector difference of two feature vectors
+
The value of (d) is 0, and the similarity characterizing the two eigenvectors is the highest. For the
In the context of 90 degrees and 180 degrees, the similarity between two eigenvectors obtained by calculating the projection length of the vector difference between the two eigenvectors on the coordinate axis in the embodiment of the present application and the similarity between the two eigenvectors obtained in the related art
The results of the similarity of the two characteristic vectors are consistent; therefore, the similarity of the two feature vectors is accurately represented based on the projection lengths of the vector difference of the two feature vectors on the X coordinate axis and the Y coordinate axis.
Continuing with the exemplary structure of the template matching device 455 provided by the embodiments of the present application implemented as software modules, in some embodiments, as shown in fig. 2, the software modules stored in the template matching device 455 of the memory 450 may include: an edge feature vector determination module 4551, configured to determine an edge feature vector of the template image; a decomposition module 4552, configured to decompose the image to be matched into a pyramid image sequence; a similarity determining module 4553, configured to perform traversal search on each pyramid image in the pyramid image sequence according to an order from top to bottom, and determine, based on a projection length of a difference between the edge feature vector and a feature vector on each pyramid image, a similarity between the edge feature vector and a feature vector on each pyramid image; a matching module 4554, configured to determine, based on the similarity between the edge feature vector and the feature vector on each pyramid image, a target image that is matched with the template image on the image to be matched.
In some embodiments, the similarity determining module 4553 is configured to slide a window on each pyramid image in the sequence of pyramid images according to an order from a top-most pyramid image to a bottom-most pyramid image, where the window is slid on a basis of a pixel point as a sliding unit.
In some embodiments, the similarity determining module 4553 is configured to perform the following operations for each pyramid image:
acquiring a characteristic vector of a region image corresponding to the window in the pyramid image;
determining a vector difference of the edge feature vector and a feature vector of the region image;
acquiring a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction;
and determining the sum of the first projection length and the second projection length, and using the sum of the first projection length and the second projection length as the similarity of the edge feature vector and the feature vector of the region image.
In some embodiments, the similarity determining module 4553 is configured to perform the following operations for each pyramid image:
determining a feature vector with the similarity smaller than a similarity threshold value with the edge feature vector as a target feature vector;
and if the number of the target feature vectors in the region image is greater than a number threshold, determining that the region image is the target image matched with the template image.
In some embodiments, the edge feature vector determination module 4551 is configured to determine gradient information of each pixel point in the template image, and determine an edge feature vector of the determined template image based on the gradient information.
In some embodiments, the edge feature vector determining module 4551 is configured to input the template image to a gradient model, and determine gradient information of each pixel point in the template image according to an output of the gradient model; the gradient information comprises a gradient direction and a gradient amplitude;
determining pixel points of which the gradient amplitudes are larger than the amplitude threshold value as edge pixel points;
and determining the feature vector of the edge pixel point as an edge feature vector.
In some embodiments, the decomposition module 4552 is configured to sample the image to be matched in a manner of gradient down-sampling, so as to obtain at least two pyramid images with different resolutions corresponding to the image to be matched; the at least two pyramid images form the pyramid image sequence.
In some embodiments, the similarity determination module 4553 is further configured to perform normalization on the edge feature vector and the feature vector on each pyramid image, respectively, so that the edge feature vector and the feature vector on each pyramid image can both be represented on a circle with a radius of 1.
Embodiments of the present application provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to execute the template matching method described in the embodiment of the present application.
Embodiments of the present application provide a computer-readable storage medium storing executable instructions, which when executed by a processor, will cause the processor to perform a method provided by embodiments of the present application, for example, a template matching method as illustrated in fig. 3 to 9.
In some embodiments, the computer-readable storage medium may be memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories.
In some embodiments, executable instructions may be written in any form of programming language (including compiled or interpreted languages), in the form of programs, software modules, scripts or code, and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.
By way of example, executable instructions may correspond, but do not necessarily have to correspond, to files in a file system, and may be stored in a portion of a file that holds other programs or data, such as in one or more scripts in a hypertext Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).
By way of example, executable instructions may be deployed to be executed on one computing device or on multiple computing devices at one site or distributed across multiple sites and interconnected by a communication network.
The above description is only an example of the present application, and is not intended to limit the scope of the present application. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present application are included in the protection scope of the present application.
Claims (10)
1. A template matching method, the method comprising:
determining an edge feature vector of the template image;
decomposing an image to be matched into a pyramid image sequence;
traversing and searching on each pyramid image in the pyramid image sequence according to the sequence from top to bottom, and respectively executing the following operations for each pyramid image: acquiring a feature vector of an area image corresponding to a sliding window in the pyramid image, determining a vector difference between the edge feature vector and the feature vector of the area image, acquiring a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction, determining a sum of the first projection length and the second projection length, and taking the sum of the first projection length and the second projection length as a similarity of the edge feature vector and the feature vector of the area image;
and determining a target image matched with the template image on the image to be matched based on each similarity.
2. The method of claim 1, wherein traversing the search over each pyramid image in the sequence of pyramid images in a top-down order comprises:
and sliding a window on each pyramid image by taking pixel points as sliding units according to the sequence from the topmost pyramid image to the bottommost pyramid image in the pyramid image sequence.
3. The method according to claim 1, wherein the determining a target image on the image to be matched that matches the template image based on each of the similarities comprises:
performing the following operations for each pyramid image:
determining the feature vector of the area image with the similarity smaller than the similarity threshold value with the edge feature vector on the pyramid image as a target feature vector;
and if the number of the target feature vectors in the region image is greater than a number threshold, determining that the region image is the target image matched with the template image.
4. The method of claim 1, wherein determining the edge feature vector of the template image comprises:
determining gradient information of each pixel point in the template image, and determining an edge feature vector of the determined template image based on the gradient information.
5. The method of claim 4, wherein determining gradient information for each pixel in the template image, and wherein determining an edge feature vector based on the gradient information comprises:
inputting the template image into a gradient model, and determining gradient information of each pixel point in the template image according to the output of the gradient model; the gradient information comprises a gradient direction and a gradient amplitude;
determining pixel points of which the gradient amplitudes are larger than the amplitude threshold value as edge pixel points;
and determining the feature vector of the edge pixel point as the edge feature vector.
6. The method of claim 1, wherein decomposing the image to be matched into a sequence of pyramid images comprises:
sampling the image to be matched in a gradient down-sampling mode to obtain at least two pyramid images with different resolutions corresponding to the image to be matched; the at least two pyramid images form the pyramid image sequence.
7. The method of claim 1, further comprising:
respectively carrying out normalization processing on the edge feature vectors and the feature vectors on the pyramid images, so that the edge feature vectors and the feature vectors on the pyramid images can be represented on a circle with the radius of 1; the feature vectors on the pyramid images comprise feature vectors of region images corresponding to sliding windows in the pyramid images.
8. A template matching apparatus, characterized in that the apparatus comprises:
the edge feature vector determining module is used for determining an edge feature vector of the template image;
the decomposition module is used for decomposing the image to be matched into a pyramid image sequence;
a similarity determining module, configured to perform traversal search on each pyramid image in the pyramid image sequence according to an order from top to bottom, and perform the following operations for each pyramid image in the pyramid image sequence, respectively: acquiring a feature vector of an area image corresponding to a sliding window in the pyramid image, determining a vector difference between the edge feature vector and the feature vector of the area image, acquiring a first projection length of the vector difference in a first direction and a second projection length of the vector difference in a second direction perpendicular to the first direction, determining a sum of the first projection length and the second projection length, and taking the sum of the first projection length and the second projection length as a similarity of the edge feature vector and the feature vector of the area image;
and the matching module is used for determining a target image matched with the template image on the image to be matched based on each similarity.
9. An electronic device, comprising:
a memory for storing executable instructions;
a processor for implementing the template matching method of any one of claims 1 to 7 when executing executable instructions stored in the memory.
10. A computer-readable storage medium having stored thereon executable instructions for, when executed by a processor, implementing the template matching method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110682462.6A CN113139626B (en) | 2021-06-21 | 2021-06-21 | Template matching method and device, electronic equipment and computer-readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110682462.6A CN113139626B (en) | 2021-06-21 | 2021-06-21 | Template matching method and device, electronic equipment and computer-readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113139626A CN113139626A (en) | 2021-07-20 |
| CN113139626B true CN113139626B (en) | 2021-10-15 |
Family
ID=76816133
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110682462.6A Active CN113139626B (en) | 2021-06-21 | 2021-06-21 | Template matching method and device, electronic equipment and computer-readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113139626B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113916897B (en) * | 2021-12-15 | 2022-03-15 | 武汉三力国创机械设备工程有限公司 | Filter element quality detection method based on image processing |
| CN116030280A (en) * | 2023-02-22 | 2023-04-28 | 青岛创新奇智科技集团股份有限公司 | Template matching method, device, storage medium and equipment |
| CN117115487B (en) * | 2023-10-23 | 2024-03-08 | 睿励科学仪器（上海）有限公司 | Template matching method, template matching system and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102157004A (en) * | 2011-04-18 | 2011-08-17 | 东华大学 | Automatic image mosaicking method for high-accuracy image measuring apparatus of super-view field part |
| CN105258681A (en) * | 2015-10-08 | 2016-01-20 | 凌云光技术集团有限责任公司 | Control for curve edge feature location and location method thereof |
| CN105608703A (en) * | 2015-12-28 | 2016-05-25 | 浙江国自机器人技术有限公司 | Current transformer oil level detection method of intelligent substation inspection robot |
| CN105825204A (en) * | 2016-04-08 | 2016-08-03 | 国家电网公司 | Method for identifying power equipment meter at night through intelligent camera |
| CN106441138A (en) * | 2016-10-12 | 2017-02-22 | 中南大学 | Deformation monitoring method based on vision measurement |
| CN107042511A (en) * | 2017-03-27 | 2017-08-15 | 国机智能科技有限公司 | The inspecting robot head method of adjustment of view-based access control model feedback |
| CN108205645A (en) * | 2016-12-20 | 2018-06-26 | 中国科学院沈阳自动化研究所 | A kind of reference map quality evaluating method of heterologous image matching system |
| CN110378376A (en) * | 2019-06-12 | 2019-10-25 | 西安交通大学 | A kind of oil filler object recognition and detection method based on machine vision |
| CN111178161A (en) * | 2019-12-12 | 2020-05-19 | 重庆邮电大学 | Vehicle tracking method and system based on FCOS |
| CN111626290A (en) * | 2019-12-31 | 2020-09-04 | 中国航天科工集团八五一一研究所 | Infrared ship target detection and identification method under complex sea surface environment |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609506B (en) * | 2008-06-20 | 2012-05-23 | 索尼株式会社 | Method and device for recognizing model object in image |
| US11195057B2 (en) * | 2014-03-18 | 2021-12-07 | Z Advanced Computing, Inc. | System and method for extremely efficient image and pattern recognition and artificial intelligence platform |
| US11074495B2 (en) * | 2013-02-28 | 2021-07-27 | Z Advanced Computing, Inc. (Zac) | System and method for extremely efficient image and pattern recognition and artificial intelligence platform |
| WO2015168777A1 (en) * | 2014-05-09 | 2015-11-12 | Fio Corporation | Discrete edge binning template matching system, method and computer readable medium |
| US20190108447A1 (en) * | 2017-11-30 | 2019-04-11 | Intel Corporation | Multifunction perceptrons in machine learning environments |
| US10755434B2 (en) * | 2018-03-28 | 2020-08-25 | Seiko Epson Corporation | Low feature object detection and pose estimation for image data streams |
| CN108537732B (en) * | 2018-04-10 | 2021-11-02 | 福州大学 | PCA-SIFT-based rapid image splicing method |
| CN108555908B (en) * | 2018-04-12 | 2020-07-28 | 同济大学 | Stacked workpiece posture recognition and pickup method based on RGBD camera |
| CN109241819A (en) * | 2018-07-07 | 2019-01-18 | 西安电子科技大学 | Based on quickly multiple dimensioned and joint template matching multiple target pedestrian detection method |
-
2021
- 2021-06-21 CN CN202110682462.6A patent/CN113139626B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102157004A (en) * | 2011-04-18 | 2011-08-17 | 东华大学 | Automatic image mosaicking method for high-accuracy image measuring apparatus of super-view field part |
| CN105258681A (en) * | 2015-10-08 | 2016-01-20 | 凌云光技术集团有限责任公司 | Control for curve edge feature location and location method thereof |
| CN105608703A (en) * | 2015-12-28 | 2016-05-25 | 浙江国自机器人技术有限公司 | Current transformer oil level detection method of intelligent substation inspection robot |
| CN105825204A (en) * | 2016-04-08 | 2016-08-03 | 国家电网公司 | Method for identifying power equipment meter at night through intelligent camera |
| CN106441138A (en) * | 2016-10-12 | 2017-02-22 | 中南大学 | Deformation monitoring method based on vision measurement |
| CN108205645A (en) * | 2016-12-20 | 2018-06-26 | 中国科学院沈阳自动化研究所 | A kind of reference map quality evaluating method of heterologous image matching system |
| CN107042511A (en) * | 2017-03-27 | 2017-08-15 | 国机智能科技有限公司 | The inspecting robot head method of adjustment of view-based access control model feedback |
| CN110378376A (en) * | 2019-06-12 | 2019-10-25 | 西安交通大学 | A kind of oil filler object recognition and detection method based on machine vision |
| CN111178161A (en) * | 2019-12-12 | 2020-05-19 | 重庆邮电大学 | Vehicle tracking method and system based on FCOS |
| CN111626290A (en) * | 2019-12-31 | 2020-09-04 | 中国航天科工集团八五一一研究所 | Infrared ship target detection and identification method under complex sea surface environment |
Non-Patent Citations (2)
| Title |
|---|
| Shape Retrieval Using Pyramid Matching with Orientation Features;Rong, Zrz等;《2009 IEEE International Conference on Intelligent Computing and Intelligent Systems》;20091228;第431-434页 * |
| 辅助动画视频分析的相似视频片段匹配技术研究;邓莉琼;《中国博士学位论文全文数据库 信息科技辑》;20140415(第4期);第I138-77页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113139626A (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113139626B (en) | Template matching method and device, electronic equipment and computer-readable storage medium | |
| US20220114759A1 (en) | Target detection method, electronic device and medium | |
| US10409856B2 (en) | Approaches for associating terms with image regions | |
| CN108256479B (en) | Face tracking method and device | |
| US8737737B1 (en) | Representing image patches for matching | |
| EP3822857B1 (en) | Target tracking method, device, electronic apparatus and storage medium | |
| US11842514B1 (en) | Determining a pose of an object from rgb-d images | |
| US11435845B2 (en) | Gesture recognition based on skeletal model vectors | |
| CN111801680A (en) | Visual feedback of process state | |
| CN109919077B (en) | Gesture recognition method, device, medium and computing equipment | |
| JP2019012526A (en) | Image processing method, computer program, and recording medium | |
| US8827710B2 (en) | Realtime user guidance for freehand drawing | |
| EP3944145A2 (en) | Method and device for training image recognition model, equipment and medium | |
| Lee et al. | Facial landmarks detection using improved active shape model on android platform | |
| Márquez-Neila et al. | Speeding-up homography estimation in mobile devices | |
| Yang et al. | Robust and real-time pose tracking for augmented reality on mobile devices | |
| CN110956131A (en) | Single-target tracking method, device and system | |
| CN112733641A (en) | Object size measuring method, device, equipment and storage medium | |
| Dubey et al. | Image alignment in pose variations of human faces by using corner detection method and its application for PIFR system | |
| Turk et al. | Computer vision for mobile augmented reality | |
| CN116089788B (en) | Online missing data processing method and device, computer equipment and storage medium | |
| Wu et al. | Evaluation of 3D LiDAR SLAM algorithms based on the KITTI dataset | |
| CN111552829A (en) | Method and apparatus for analyzing image material | |
| Pereira et al. | Mirar: Mobile image recognition based augmented reality framework | |
| Holzer et al. | Efficient learning of linear predictors for template tracking |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 310051 8 / F, building a, 1181 Bin'an Road, Binjiang District, Hangzhou City, Zhejiang Province Applicant after: Zhejiang Huarui Technology Co.,Ltd. Address before: 310051 8 / F, building a, 1181 Bin'an Road, Binjiang District, Hangzhou City, Zhejiang Province Applicant before: ZHEJIANG HUARAY TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |