A visual positioning algorithm applied to industrial production line

Abstract

The application discloses an industrial production line visual positioning algorithm, and has the characteristics that an industrial camera and a software system are arranged; an image is collected in real time by using the industrial camera, the collected image is transmitted to an image processing system, and a corner point feature in the image, i.e., a corner point feature of a target image, is extracted by using a detection algorithm in the image processing system; the extracted corner point feature of the target image is matched with a corner point feature in a known matching template, the matching result obtained is screened out, and a final matching mode is obtained; an actual production line is tested, parameters of a visual positioning algorithm are adjusted according to a test result, a corner point feature in an image is extracted by using a corner point detection algorithm, and matching is performed in combination with a geometric distribution law between template feature points, so that the positioning precision is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of machine vision technology, and in particular to a visual positioning algorithm applied to industrial production lines. Background Technology

[0002] Visual positioning refers to the technology of using cameras or other visual sensors to acquire environmental information and determine the position and orientation of objects in space. On industrial production lines, visual positioning technology enables the precise positioning and identification of different objects. For example, on assembly lines, robots can accurately grasp and place parts using visual positioning technology. This not only improves production efficiency but also reduces the error rate of manual operations.

[0003] Product positioning is the prerequisite and foundation for target recognition, defect detection, size measurement, and grasping and placement. Accurately identifying the position and orientation of a product in a visual positioning and inspection system is the core technology of the system. Visual positioning technology involves a large amount of image processing and feature matching operations, placing high demands on computing resources. How to improve computational efficiency while ensuring positioning accuracy is a pressing problem to be solved in practical applications.

[0004] Visual localization is based on image information captured by a camera, and determines the position and pose of an object by processing and analyzing the image. Core technologies in visual localization include feature extraction and feature matching. Feature extraction refers to extracting feature points or feature descriptors related to the object's position and pose from an image. Commonly used feature extraction algorithms include Harris corner detection, SIFT, and SURF, which help identify key feature points in an image. Feature matching involves matching the extracted feature points in the image with known templates or maps to determine the object's position in space. Matching algorithms typically include KNN matching and RANSAC, which achieve matching by calculating the similarity between feature points. Summary of the Invention

[0005] This application provides a visual positioning algorithm for industrial production lines. It uses a corner detection algorithm to extract corner features from images and combines them with the geometric distribution patterns between template feature points for matching, which effectively improves the positioning accuracy. By calculating the positional relationship between corner points, it obtains feature information with rotation invariant properties, enabling the algorithm to maintain high-precision positioning capability when dealing with product rotation.

[0006] This application provides a visual positioning algorithm for industrial production lines, including:

[0007] S101, equipped with industrial cameras and software systems;

[0008] S102 uses an industrial camera to acquire images in real time, transmits the acquired images to an image processing system, and uses the detection algorithm in the image processing system to extract corner features in the image, i.e., corner features of the target image.

[0009] S103, the corner features of the extracted target image are matched with the corner features in the known matching template, and the matching results are filtered and eliminated to obtain the final matching method.

[0010] S104. Test the actual production line and adjust the parameters of the visual positioning algorithm based on the test results.

[0011] Preferably, the geometric distribution pattern between the corner feature points of the matching template is used as the matching basis, and a matching strategy of pairwise grouping of corner points is adopted. If the template contains n corner points, then a matching pattern can be formed. Given a combination of corner points, where each corner point contains information represented by (x, y, θ) indicating its coordinates and angle respectively, each combination contains two corner point positions (x1, y1, θ1) and (x2, y2, θ2). These positions are obtained by calculating the positional relationship between the two corner points. Δθ1=θ1-θ,Δθ2=θ2-θ,where d represents the distance between the two corner points, and Δθ1 and Δθ2 represent the angular difference between the two corner points relative to a certain reference direction. d, Δθ1, and Δθ2 are information with rotational invariance.

[0012] Preferably, the matching template contains n corner points, namely pS[0], pS[1]...pS[n-1], and the target image to be located currently detects m corner points, namely pD[0], pD[1]...pD[m-1]. Since the target product is a rigid body, the relative positional relationship between any two corner points pS[i] and pS[j] is fixed and has translational and rotational invariance, so that they satisfy the same relative positional relationship within the preset tolerance range. The formula is used. Calculate the distance between pS[i] and pS[j] in the matching template, where d ji This represents the distance between pS[i] and pS[j], (x i y i (x) represents the coordinates of corner point pS[i], (x) j y j () represents the coordinates of pS[j]. Find the two corner points pD[I] and pD[J] from pD. Similarly, use the formula... Calculate the distance between pD[I] and pD[J] in the target image, where d JI This represents the distance between pD[I] and pD[J], (x I yI (x) represents the coordinates of corner point pD[I], (x) J y J ) represents the coordinates of pD[J].

[0013] Preferably, if m corner points are detected in the target image, then after grouping, a sequence is formed. A combination of these, for the target image In each combination and template Feature matching requires a total of [number] matches. Second-rate.

[0014] Preferably, the difference Δd = |d| between the corner distances in the target image and the template image is calculated. JI -d ji The calculated distance difference is compared with the preset tolerance range. If the distance difference between corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) is less than the preset tolerance range, the corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) are considered to satisfy the matching relationship; otherwise, they are not satisfied.

[0015] Preferably, when the matching relationship is satisfied, the template corner point pS[i] corresponds to the corner point pD[I] of the target image, and the template corner point pS[j] corresponds to the corner point pD[J] of the target image. The algorithm uses a voting method to increment the two matching methods by 1 respectively, and so on, to perform voting statistics on all matching methods. After filtering and eliminating the statistical results, the one with the most votes is the best matching method, and the number of votes for each possible matching method is initialized to 0.

[0016] Preferably, the screening and elimination method is as follows:

[0017] 1) The votes received by the candidate with the fewest votes will be considered invalid, cleared to zero, and discarded.

[0018] 2) For those who are excluded, their votes are invalidated, and the number of votes received by those who voted for them is reduced by 1 accordingly;

[0019] 3) Repeat steps 1 and 2 until the votes received by the person with the fewest votes are equal to the votes received by the person with the most votes;

[0020] 4) At this point, each contestant has the same number of votes, and the total number of votes equals the number of remaining contestants minus 1, thus ending the cycle;

[0021] Through the above operations, any two remaining contestants vote for each other, which is equivalent to mutual verification between any two matching methods (pS[i], pD[I]) and (pS[j], pD[J]) in the remaining matching methods, where i,j∈[0,n), I,J∈[0,m), and the corner point group pD[I] in the target image, where I∈[0,m), is the final matching result.

[0022] One or more technical solutions provided in this application have at least the following technical effects or advantages: The Harris corner detection algorithm is used to extract corner features from images, and matching is performed based on the geometric distribution patterns between template feature points, effectively improving positioning accuracy. By calculating the positional relationships between corner points, rotation-invariant feature information is obtained, enabling the algorithm to maintain high-precision positioning capabilities even when dealing with product rotation. The algorithm design considers computational efficiency; by optimizing parameters such as the corner detection threshold and matching tolerance, unnecessary computation is reduced, and processing speed is improved. Testing on different hardware platforms ensures the algorithm's performance stability in various environments, meeting the real-time requirements of industrial production lines. Algorithm parameters (such as corner detection threshold and matching tolerance) can be adjusted based on actual test results to optimize positioning performance, making the algorithm more flexible and practical. Through performance evaluation and testing, problems and shortcomings in the algorithm can be identified in a timely manner, providing direction for subsequent improvements and optimizations. Attached Figure Description

[0023] Figure 1 This is a schematic diagram illustrating the features of a visual positioning algorithm applied to an industrial production line according to the present invention.

[0024] Figure 2 This is an experimental result diagram of corner feature extraction according to an embodiment of the present invention;

[0025] Figure 3 This is an experimental result diagram of corner feature matching according to an embodiment of the present invention. Detailed Implementation

[0026] To facilitate understanding of the present invention, a more complete description of this application will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the present invention.

[0027] It should be noted that the terms "vertical," "horizontal," "up," "down," "left," "right," and similar expressions used in this article are for illustrative purposes only and do not represent the only possible implementation.

[0028] Unless otherwise defined, 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 invention pertains; the terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention; the term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] Example 1: Figure 1 This is a flowchart illustrating a visual positioning algorithm applied to an industrial production line according to an embodiment of the present invention, including:

[0030] S101, equipped with industrial cameras and software systems;

[0031] Specifically, the Baslerace 2500 series industrial camera, known for its high resolution, high frame rate, and excellent color reproduction, is selected. Based on the product type and material, LEDs are chosen as the lighting source. LEDs provide uniform and adjustable illumination, reducing glare. The light source layout is designed according to the product's shape, size, and shooting angle. This layout ensures even illumination across the entire product surface while avoiding direct light from the camera lens to prevent glare or shadows that could affect image quality. Different types of light source layouts, such as ring lights, strip lights, or coaxial lights, are employed, selected based on the specific needs of the product. By adjusting the position, angle, and brightness of the light sources, the lighting effect can be optimized, reducing shadows and reflections. Impact on image quality; Select camera brackets and fixing devices according to the production line layout and camera shooting angle, ensuring that the camera brackets and fixing devices are installed firmly and reliably to withstand vibration and shaking on the production line. Use fasteners and connectors to ensure that the connection between the camera and the bracket is tight and stable. Adjust the camera angle and focal length according to the product positioning requirements and shooting range. Use the camera's built-in adjustment mechanism or external adjustment tools to ensure that the camera can accurately aim at the product and capture a clear image. During the adjustment process, pay attention to the camera's field of view and depth of field to ensure that the captured image meets the requirements of subsequent processing.

[0032] Based on the production line and visual positioning algorithm requirements, an integrated software system incorporating the visual positioning algorithm is developed. This system includes functions for image reception, processing, analysis, and positioning result output. Parameter adjustments and performance optimizations are performed on the algorithm according to the actual production line conditions and algorithm performance requirements. The stability and accuracy of the algorithm within the software system are verified to ensure reliable positioning results.

[0033] S102 uses an industrial camera to acquire images in real time, transmits the acquired images to an image processing system, and uses the detection algorithm in the image processing system to extract corner features in the image, i.e., corner features of the target image.

[0034] Furthermore, industrial cameras are used to ensure they are functioning properly and capturing images in real time. The resolution and frame rate of the industrial cameras are set to meet the needs of subsequent image processing and positioning. The industrial cameras capture product images on the production line in real time and transmit these images to the image processing system for further processing. The captured images should be clear, unblurred, and contain sufficient product details to extract effective feature information.

[0035] In image processing systems, the Harris corner detection algorithm is used to extract corner features from captured images. While a mature and stable algorithm capable of quickly and effectively extracting corner features, the Harris corner detection algorithm acquires limited information, containing only position and angle information for each corner. In scenarios where product rotation is possible, it cannot determine the one-to-one correspondence between template corners and image corners. Addressing the matching problem of the Harris corner detection algorithm in rotating scenarios, this embodiment utilizes the geometric distribution patterns between template feature points as a matching basis. It searches for corners with similar distribution patterns in the image, employing a pairwise grouping matching strategy. Assuming the template contains n corners, then... There are several combinations. Assuming each corner point contains information represented by (x, y, θ) indicating its coordinates and angle respectively, then each combination of corner points contains two corner point positions: (x1, y1, θ1) and (x2, y2, θ2). By calculating the positional relationship between the two corner points, the following rotationally invariant feature information can be obtained: Δθ1=θ1-θ,Δθ2=θ2-θ,where d represents the distance between the two corner points, and Δθ1 and Δθ2 represent the angular difference between the two corner points relative to a certain reference direction. d, Δθ1, and Δθ2 are information with rotational invariance.

[0036] A specific example is as follows: Images of parts on the production line are captured in real time using an industrial camera and transmitted to an image processing system for further processing. The industrial camera is set to a resolution of 1920x1080 and a frame rate of 30fps to ensure that the captured images are clear, blur-free, and contain sufficient part details. In a single part image, the Harris corner detection algorithm extracts five corner points, labeled A, B, C, D, and E. The corner point combination (A, B) is selected for detailed analysis. The information for corner point A is (x1, y1, θ1), with specific values ​​of (100, 150, 30°). The information for corner point B is (x2, y2, θ2), with specific values ​​of (200, 250, 60°). The positional relationship between the two corner points is calculated to obtain rotationally invariant feature information.

[0037] Δθ1=θ1-θ A , where θ A Let θ be the angle between the direction of the line connecting A and B and the x-axis. A =45°, then Δθ1=60°-45°=15°; Δθ2=θ2-θ B Similarly, let θ B Let θ be the angle of point B relative to a certain global reference direction. B =75°, then Δθ2 = 30° - 75° = -45°.

[0038] S103, the corner features of the extracted target image are matched with the corner features in the known matching template, and the matching results are filtered and eliminated to obtain the final matching method.

[0039] Specifically, such as Figure 3 As shown, if m corner points are detected in the target image, then after grouping, they form... A combination of these, for the target image In each combination and template Feature matching requires a total of [number] matches. Next, suppose the matching template contains n corner points, namely pS[0], pS[1]...pS[n-1], and the target image to be located currently detects m corner points, namely pD[0], pD[1]...pD[m-1]. Since the target product is a rigid body, the relative positional relationship between any two corner points pS[i] and pS[j] is fixed and has translational and rotational invariance, so that they satisfy the same relative positional relationship within the preset tolerance range, such as... Figure 2 Using formula Calculate the distance between pS[i] and pS[j] in the matching template, where d ji This represents the distance between pS[i] and pS[j], (x i y i (x) represents the coordinates of corner point pS[i], (x) j y j () represents the coordinates of pS[j]. Find the two corner points pD[I] and pD[J] from pD. Similarly, use the formula... Calculate the distance between pD[I] and pD[J] in the target image, where d JI This represents the distance between pD[I] and pD[J], (x I y I (x) represents the coordinates of corner point pD[I], (x) J y J () represents the coordinates of pD[J], and the difference Δd = |d| between the corner distances in the target image and the template image is calculated. JI -dji The calculated distance difference is compared with a preset tolerance range. When the distance difference between corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) is less than the preset tolerance range, the corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) are considered to satisfy the matching relationship; otherwise, they are not. When the matching relationship is satisfied, the template corner point pS[i] corresponds to the corner point pD[I] of the target image, and the template corner point pS[j] corresponds to the corner point pD[I] of the target image. The corner point pD[J] of the target image, i.e., the matching method (pS[i], pD[i]) and the matching method (pS[j], pD[J]) corroborate each other. The algorithm uses a voting method to increment the number of votes for each matching method by 1, and so on, to count the votes for all matching methods. After filtering and eliminating the statistical results, the one with the most votes is the best matching method. The number of votes for each possible matching method (i.e., the correspondence between the template corner point and the target image corner point) is initialized to 0. The specific filtering and elimination method is as follows:

[0040] 1. The votes received by the candidate with the fewest votes are considered invalid, are cleared to zero, and are discarded.

[0041] After the voting ends, the number of votes for each matching method is counted, and the matching method with the fewest votes is identified. The matching method with the fewest votes means that it has the lowest support among all possible matching methods, so it is considered unreliable, its vote count is reset to zero, and it is removed from the candidate matching methods.

[0042] 2. For those who are excluded, their votes are invalidated, and the number of votes received by those who voted for them is reduced by 1 accordingly;

[0043] Since the eliminated match has a corroborating relationship with other match types (i.e., it voted for other match types), its votes need to be considered invalid. At the same time, the votes of those match types that accepted the votes of the eliminated match type need to be reduced by 1 to reflect this change.

[0044] 3. Repeat steps 1 and 2 until the votes received by the person with the fewest votes are equal to the votes received by the person with the most votes;

[0045] Repeat the above filtering and elimination process until the number of votes for the least popular match is the same as the number of votes for the most popular match. This means that the remaining matches have reached a certain balance in terms of support, and no further obviously unreliable matches can be eliminated.

[0046] 4. At this point, each contestant has the same number of votes, and the total number of votes equals the number of remaining contestants minus 1, thus ending the cycle;

[0047] When the loop terminates, the remaining matches should have the same number of votes, and this number of votes should be equal to the number of remaining matches minus 1. This is because, ideally, each remaining match should be corroborated by all other remaining matches (i.e., each match votes for all other matches), so their vote counts should be equal.

[0048] Through the above screening and elimination process, we obtained the final set of matching methods. In this set, any two matching methods corroborate each other, that is, the feature differences between them are within the tolerance range, and their support (votes) are consistent. Therefore, these matching methods are reliable. The corner point group pD[I] (where I∈[0,m)) in the target image is the final result of matching the template corner point. Through the above operation, any two of the remaining candidates vote for each other. This is equivalent to any two matching methods (pS[i], pD[I]) and (pS[j], pD[J]) corroborating each other in the remaining matching methods, where i,j∈[0,n), I,J∈[0,m). Through the above method, we know that the corner point group pD[I] in the target image, where I∈[0,m), is the final matching result.

[0049] S104, Test on the actual production line, and adjust the parameters of the visual positioning algorithm based on the test results;

[0050] Furthermore, the positioning accuracy is evaluated by comparing the algorithm's positioning results with the actual position of the product. Statistical metrics such as mean squared error (MSE) and mean absolute error (MAE) are used to quantify the positioning accuracy. Tests are conducted under different lighting conditions, product poses, and background environments. The time required for the algorithm to process a single image is measured to evaluate its computational efficiency. Tests are also performed on different hardware platforms (such as CPU and GPU) to compare the algorithm's performance in different environments. Adjusting the corner detection threshold can control the number and quality of detected corners. Too many or too few detected corners will lead to a decrease in positioning accuracy. Therefore, the corner detection threshold needs to be adjusted based on actual test results to achieve the best effect. In the feature matching process, the allowable tolerance directly affects the accuracy of the matching. If the allowable tolerance is set too large, it may lead to false matches; if the allowable tolerance is set too small, it may lead to matching failure. Therefore, the matching tolerance needs to be adjusted based on actual test results to achieve the best matching effect.

[0051] In practical applications, poor image quality, product deformation, and other anomalies may cause the algorithm to fail to locate or produce significant errors. When image quality is poor (e.g., blurry, noisy), preprocessing techniques (such as filtering and enhancement) can be used to improve image quality. If preprocessing still cannot meet the positioning requirements, feature extraction and matching algorithms can be used. Product deformation may cause changes in the position of feature points, thus affecting positioning accuracy. A deformation model can be introduced to estimate the product deformation and correct the position of feature points. An anomaly monitoring mechanism should be designed to monitor the algorithm's running status in real time, promptly detect and handle anomalies, and trigger an alarm mechanism when an anomaly is detected to notify operators to intervene or take appropriate remedial measures.

[0052] The technical solutions described in the embodiments of this application have at least the following technical effects or advantages: The Harris corner detection algorithm is used to extract corner features from images, and the geometric distribution patterns between template feature points are matched, effectively improving positioning accuracy. By calculating the positional relationships between corner points, feature information with rotational invariance is obtained, enabling the algorithm to maintain high-precision positioning capability even when dealing with product rotation. The algorithm design considers computational efficiency; by optimizing parameters such as the corner detection threshold and matching tolerance, unnecessary computation is reduced, and processing speed is improved. Testing on different hardware platforms ensures the performance stability of the algorithm under various environments, meeting the real-time requirements of industrial production lines. Algorithm parameters (such as corner detection threshold and matching tolerance) can be adjusted according to actual test results to optimize positioning performance, making the algorithm more flexible and practical. Through performance evaluation and testing, problems and deficiencies in the algorithm can be identified in a timely manner, providing direction for subsequent improvements and optimizations.

[0053] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A visual positioning algorithm applied to industrial production lines, characterized in that, include: S101, equipped with industrial cameras and software systems; S102 uses an industrial camera to acquire images in real time, transmits the acquired images to an image processing system, and uses the detection algorithm in the image processing system to extract corner features in the image, i.e., corner features of the target image. S103, the extracted corner features of the target image are matched with the corner features in the known matching template. The geometric distribution of corner feature points in the matching template is used as the matching basis. A pairwise matching strategy is adopted, grouping corner points in pairs. Based on the distance between corner points pS[i] and pS[j] in the matching template and the distance between corner points pD[i] and pD[j] in the target image, the difference Δd between the corner distances in the target image and the template image is calculated. The calculated distance difference is compared with the preset tolerance range. When the distance difference between corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) is less than the preset tolerance range, it is considered that corner point pair (pS[i], pS[j]) and corner point pair (pD[I], pD[J]) satisfy the matching relationship; otherwise, it does not satisfy the matching relationship. When the matching relationship is satisfied, the template corner point pS[i] corresponds to the corner point pD[I] of the target image, and the template corner point pS[j] corresponds to the corner point pD[J] of the target image. The algorithm uses a voting method to increment the number of votes for each of the two matching methods by 1, and so on. The voting statistics are performed on all matching methods. After filtering and eliminating the statistical results, the one with the most votes is the best matching method. The number of votes for each possible matching method is initialized to 0. The final matching result is obtained by eliminating the matching method with the fewest votes and invalidating its votes. The votes of the matching methods that received its votes are reduced by 1 accordingly. This step is repeated until the number of votes of the one with the fewest votes is equal to the number of votes of the one with the most votes. S104. Test the actual production line and adjust the parameters of the visual positioning algorithm based on the test results.

2. The visual positioning algorithm for industrial production lines as described in claim 1, characterized in that, If a template contains n corner points, it can form Such combinations, each corner point contains information in the form of ( , θ and y represent the coordinates and angle of a corner point, respectively. Therefore, each combination of corner points contains two corner point positions (...). , , )and( , , This is obtained by calculating the positional relationship between two corner points: , Δθ1=θ1-θ, Δθ2=θ2-θ, where d represents the distance between the two corner points, Δθ1 and Δθ2 represent the angular difference between the two corner points relative to a certain reference direction, and d, Δθ1, and Δθ2 are information with rotational invariance.

3. The visual positioning algorithm for industrial production lines as described in claim 1, characterized in that, The matching template contains n corner points, namely pS[0], pS[1]...pS[n-1]. The target image to be located currently has m corner points detected, namely pD[0], pD[1]...pD[m-1]. Since the target product is a rigid body, the relative positional relationship between any two corner points pS[i] and pS[j] is fixed and has translational and rotational invariance, so that they satisfy the same relative positional relationship within the preset tolerance range. The formula is used. Calculate the distance between corner point pS[i] and corner point pS[j] in the matching template, where, This represents the distance between pS[i] and pS[j]. , ) represents the coordinates of corner point pS[i], ( , () represents the coordinates of pS[j]. Find the two corner points pD[I] and pD[J] from pD. Similarly, use the formula Calculate the distance between corner point pD[I] and corner point pD[J] in the target image, where, This represents the distance between pD[I] and pD[J]. , ) represents the coordinates of corner point pD[I], ( , ) represents the coordinates of pD[J].

4. The visual positioning algorithm for industrial production lines as described in claim 1, characterized in that, If m corner points are detected in the target image, then after grouping, they form... A combination of these, for the target image In each combination and template Feature matching requires a total of [number] matches. Second-rate.

5. The visual positioning algorithm for industrial production lines as described in claim 1, characterized in that, The specific filtering and elimination methods are as follows: 1) The votes received by the candidate with the fewest votes will be considered invalid, cleared to zero, and discarded. 2) For those who are excluded, their votes are invalidated, and the number of votes received by those who voted for them is reduced by 1 accordingly; 3) Repeat steps 1 and 2 until the votes received by the person with the fewest votes are equal to the votes received by the person with the most votes; 4) At this point, each contestant has the same number of votes, and the total number of votes equals the number of remaining contestants minus 1, thus ending the cycle; Through the above operations, any two remaining contestants vote for each other, which is equivalent to mutual verification between any two matching methods (pS[i], pD[I]) and (pS[j], pD[J]) in the remaining matching methods, where i,j∈[0,n), I,J∈[0,m), and the corner point group pD[I] in the target image, where I∈[0,m), is the final matching result.

Images