Image feature point enhancement method and device, computer equipment and storage medium

Abstract

The invention relates to an image feature point enhancement method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring an original image; preprocessing the original image to obtain a first image; according to a Hessian matrix, obtaining a first feature point set from the first image; performing non-maximum suppression screening on the first feature point set according to an MASK grid to obtain a second feature point set; and according to a Feature Booster algorithm, the second feature point set is enhanced, so that an enhanced feature point set can be obtained. According to the method, stable feature points can be obtained, and the feature points can be enhanced.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular, to an image feature point enhancement method, apparatus, computer device, and storage medium. Background Art

[0002] With the development of information technology and the Internet, there is a higher demand for humanoid robots.

[0003] Image processing such as image recognition, map recognition, position recognition, and real-time map drawing is the basis for the development of humanoid robots. In the existing image processing methods, the accuracy and precision are relatively high when the texture is clear and the lighting is fixed. However, in scenarios where the texture of the image is unclear and the lighting changes frequently, the accuracy and precision are relatively low, making it difficult to meet the requirements of humanoid robots.

[0004] It can be seen that the existing image processing methods are still difficult to meet the requirements. Summary of the Invention

[0005] To solve the above technical problems or at least partially solve the above technical problems, the present invention provides an image feature point enhancement method, apparatus, computer device, and storage medium.

[0006] In a first aspect, the present invention provides an image feature point enhancement method, the method comprising:

[0007] Obtain an original image;

[0008] Preprocess the original image to obtain a first image;

[0009] According to the Hessian matrix, obtain a first feature point set from the first image;

[0010] Perform non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set;

[0011] Enhance the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set.

[0012] Optionally, the preprocessing the original image to obtain a first image includes:

[0013] Detect whether the original image is legal;

[0014] If the original image is illegal, output an empty set;

[0015] If the original image is legal, filter the original image to obtain the first image.

[0016] Optionally, each first feature point in the first feature point set includes: a first key point and a corresponding BRIEF descriptor;

[0017] The obtaining of the first feature point set from the first image according to the Hessian matrix includes:

[0018] Obtaining a first key point set from the first image;

[0019] Extracting the BRIEF descriptor of each first key point from the first image and the first key point set through the BRIEF algorithm.

[0020] Optionally, the key point includes an attribute, and the obtaining of the first key point set from the first image includes:

[0021] Obtaining a set of original feature points from the first image, where the set of original feature points includes a plurality of original feature points;

[0022] Obtaining the Hessian matrix of each original feature point;

[0023] Obtaining the characteristic response value of the Hessian matrix of the original feature point at multiple scale levels;

[0024] Taking the original feature points whose corresponding characteristic response values are greater than or equal to a preset threshold as first candidate feature points;

[0025] Screening the first candidate feature points according to the corner point enhancement algorithm, and taking the set of the screened first candidate feature points as a second candidate feature point set;

[0026] Obtaining the attribute of each second candidate feature point in the second candidate feature point set, and taking the second candidate feature point set carrying the attribute as the first key point set.

[0027] Optionally, the attribute includes a key point direction, and the obtaining of the attribute of each second candidate feature point in the second candidate feature point set includes:

[0028] Obtaining the centroid of the preset neighborhood of the second candidate feature point;

[0029] Taking the direction from the second candidate feature point to the centroid as the key point direction.

[0030] Optionally, the non-maximum suppression screening of the first feature point set according to the MASK grid to obtain a second feature point set includes:

[0031] Performing non-maximum suppression screening on one first feature point in the first feature point set to obtain one second feature point;

[0032] Among them, all the second feature points constitute the second feature point set.

[0033] Optionally, the non-maximum suppression screening of one of the first feature points in the first feature point set to obtain a second feature point includes:

[0034] Obtaining a set of all first feature points within a preset range around the first feature point as a set of candidate feature points;

[0035] Obtaining the feature response value of each of the first feature points in the set of candidate feature points;

[0036] Taking the first feature point with the maximum feature response value as the second feature point.

[0037] Optionally, the enhancement of the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set includes:

[0038] Denosing the second feature point set to obtain a denoised point set;

[0039] Obtaining a multi-scale point set from the denoised point set at multiple scale levels;

[0040] Screening the multi-scale points in the multi-scale point set according to the corner enhancement algorithm to obtain a corner point set;

[0041] After rotating and normalizing the key point directions of each corner point in the corner point set, the enhanced feature point set is obtained.

[0042] In a second aspect, an image feature point enhancement device is disclosed, and the device includes:

[0043] An image acquisition unit for acquiring an original image;

[0044] An image processing unit for preprocessing the original image to obtain a first image;

[0045] A feature point acquisition unit for acquiring a first feature point set from the first image according to the Hessian matrix;

[0046] The feature point acquisition unit is further configured to perform non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set;

[0047] The feature point acquisition unit is further configured to enhance the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set.

[0048] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the method described in any one of the above is implemented.

[0049] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in any one of the above is implemented.

[0050] The present invention provides an image feature point enhancement method, apparatus, computer device, and storage medium. The method includes: obtaining an original image; preprocessing the original image to obtain a first image; obtaining a first feature point set from the first image according to the Hessian matrix; performing non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set; and enhancing the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set. The image feature point enhancement method according to the embodiments of the present invention uses the Hessian matrix for feature point extraction, which can improve the robustness of feature point detection and the accuracy of descriptors. And Feature Booster can make the extracted features more robust, have better discrimination, and be more stable. Therefore, the method according to the embodiments of the present invention can improve the stability of the extracted feature points, enhance the feature points, and avoid mis-matching of subsequent feature points. The method according to the embodiments of the present invention has outstanding effects especially in application scenarios where feature points are unclear and unstable due to light changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present invention and used together with the specification to explain the principles of the present invention.

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0053] Figure 1 The following shows an application environment diagram of the image feature point enhancement method according to the embodiments of the present invention;

[0054] Figure 2 The following shows a flowchart of the image feature point enhancement method according to the embodiments of the present invention;

[0055] Figure 3 The following shows a structural block diagram of the image feature point enhancement method apparatus according to the embodiments of the present invention;

[0056] Figure 4 The internal structure diagram of the computer device in the embodiment of the present invention is shown as follows. Specific embodiments

[0057] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0058] Figure 1 It is an application environment diagram of the image feature point enhancement method in an embodiment. Refer to Figure 1 , the image feature point enhancement method is applied to an image feature point enhancement system. The image feature point enhancement method includes a terminal 110 and / or a server 120. The terminal 110 and the server 120 are connected through a network. The terminal 110 may specifically be a desktop terminal or a mobile terminal, and the mobile terminal may specifically be at least one of a mobile phone, a tablet computer, a notebook computer, etc. The server 120 may be implemented by an independent server or a server cluster composed of multiple servers.

[0059] The image feature point enhancement method of the present invention is applied to the terminal 110 and / or the server 120.

[0060] Figure 2 The flowchart of the image feature point enhancement method in the embodiment of the present invention is shown as follows. As Figure 2 shown, the method includes:

[0061] Step 210, obtaining an original image;

[0062] Step 220, preprocessing the original image to obtain a first image;

[0063] Step 230, obtaining a first feature point set from the first image according to the Hessian matrix;

[0064] Step 240, performing non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set;

[0065] Step 250, enhancing the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set.

[0066] In the embodiments of the present invention, the original image can be an RGB image or a depth image captured by a humanoid robot. The original image can be used for Simultaneous Localization and Mapping (SLAM). SLAM can include 2D SLAM, 3D SLAM, and Visual SLAM. Whichever type of SLAM it is, it is based on feature points extracted from the image.

[0067] In the image feature point enhancement method of the embodiments of the present invention, using the Hessian matrix for feature point extraction can improve the robustness of feature point detection and the accuracy of descriptors. And Feature Booster can make the extracted features more robust, have better distinctiveness, and be more stable. Therefore, the method of the embodiments of the present invention can improve the stability of the extracted feature points, enhance the feature points, and avoid mis-matching of subsequent feature points. The method of the embodiments of the present invention has outstanding effects especially in application scenarios where feature points are unclear and unstable due to light changes.

[0068] In the embodiments of the present invention, in step 220, the preprocessing of the original image to obtain the first image includes:

[0069] Detect whether the original image is legal;

[0070] If the original image is not legal, output an empty set;

[0071] If the original image is legal, filter the original image to obtain the first image.

[0072] In the embodiments of the present invention, filtering the image can be Gaussian filtering, median filtering, or other filtering methods, which can remove noise in the image and reduce the influence of noise on feature point detection and descriptors.

[0073] In the embodiments of the present invention, each first feature point in the first feature point set includes: a first key point and a corresponding BRIEF descriptor;

[0074] In step 230, obtaining the first feature point set from the first image according to the Hessian matrix includes:

[0075] Obtain a first key point set from the first image;

[0076] Extract the BRIEF descriptor of each first key point from the first image and the first key point set through the BRIEF algorithm.

[0077] In the embodiments of the present invention, the key point includes attributes, and obtaining the first key point set from the first image includes:

[0078] Obtain a set of original feature points from the first image, where the set of original feature points includes a plurality of original feature points;

[0079] Obtain the Hessian matrix of each of the original feature points;

[0080] At multiple scale levels, obtain the feature response values of the Hessian matrix of the original feature points;

[0081] Take the original feature points for which the corresponding feature response values are greater than or equal to a preset threshold as the first candidate feature points;

[0082] According to the corner enhancement algorithm, screen the first candidate feature points, and take the set of the screened first candidate feature points as the second candidate feature point set;

[0083] Obtain the attributes of each second candidate feature point in the second candidate feature point set, and take the second candidate feature point set carrying the attributes as the first key point set.

[0084] In an embodiment of the present invention, the attribute includes the key point direction, and obtaining the attributes of each second candidate feature point in the second candidate feature point set includes:

[0085] Obtain the centroid of a preset neighborhood of the second candidate feature point;

[0086] Take the direction from the second candidate feature point to the centroid as the key point direction.

[0087] The Hessian matrix, also translated as the Hessian matrix, Hesse matrix, etc., is a square matrix composed of the second-order partial derivatives of a multivariate function, which describes the local curvature of the function.

[0088] In an embodiment of the present invention, at multiple scale levels, obtaining the feature response values of the Hessian matrix of the original feature points can be through a Gaussian pyramid. The Gaussian pyramid is from bottom to top, and the nodes included in the upper layer are a preset ratio of the nodes included in the lower layer. For example, the first layer from bottom to top includes 1000 nodes, and the second layer includes 500 nodes; or the first layer includes 1000 nodes, and the second layer includes 250 nodes. In an embodiment of the present invention, the feature response values of the Hessian matrix can be calculated in the first layer and the feature response values of the Hessian matrix can be calculated in the second layer.

[0089] The corner enhancement algorithm is actually a corner detection method based on the pixel brightness change, and determines whether a point is a corner through the pixel brightness change within a preset area.

[0090] In an embodiment of the present invention, according to the corner enhancement algorithm, the first candidate feature points are screened. Specifically, for the points within a preset range around a certain first candidate feature point, a value is calculated according to the corner enhancement algorithm. If it is greater than the preset threshold, the point is retained; if it is less than or equal to the preset threshold, the point is screened out. The preset range in the corner enhancement algorithm can be a circular area, such as a 5*5 circular area.

[0091] In an embodiment of the present invention, the feature points include key points and BRIEF descriptors, and the key points include attributes, which can be position / coordinates and direction.

[0092] When obtaining the centroid of a certain second candidate feature point, the preset domain refers to all the second candidate feature points within a preset range centered on a certain second candidate feature point. The preset range can be 10*10, or 8*8, or other ranges, which will not be elaborated here.

[0093] At multiple scales, obtaining the feature response values of the Hessian matrix of the original feature points can enhance the adaptability of the feature points to scale changes, so as to finally obtain feature points that are stable at multiple scales, thereby improving the accuracy and precision.

[0094] In an embodiment of the present invention, in step 240, the non-maximum suppression screening of the first feature point set according to the MASK grid to obtain the second feature point set includes:

[0095] Performing non-maximum suppression screening on a first feature point in the first feature point set to obtain a second feature point;

[0096] Among them, all the second feature points constitute the second feature point set.

[0097] In an embodiment of the present invention, the performing non-maximum suppression screening on a first feature point in the first feature point set to obtain a second feature point includes:

[0098] Obtaining the set of all first feature points within a preset range around the first feature point as the set of candidate feature points;

[0099] Obtaining the feature response value of each first feature point in the set of candidate feature points;

[0100] Taking the first feature point with the largest feature response value as the second feature point.

[0101] In an embodiment of the present invention, in step 250, the enhancement of the second feature point set according to the Feature Booster algorithm to obtain the enhanced feature point set includes:

[0102] Denoise the second feature point set to obtain a denoised point set;

[0103] On multiple scale levels, obtain a multi-scale point set from the denoised point set;

[0104] According to the corner enhancement algorithm, screen the multi-scale points in the multi-scale point set to obtain a corner point set;

[0105] After rotating and normalizing the key point directions of each corner point in the corner point set, obtain the enhanced feature point set.

[0106] In the embodiments of the present invention, denoising, multi-scale levels, and rotation normalization of key point directions can all improve the stability of the enhanced feature points in the finally obtained enhanced feature point set, so that when the light changes, the extracted feature points are stable, making the subsequent SLAM and other processing more accurate and precise.

[0107] The embodiments of the present invention can enhance feature points, especially improve the recognition rate of feature points in application scenarios with light changes, and avoid false matching.

[0108] As Figure 3 shown, the present invention also provides an image feature point enhancement device, and the device includes:

[0109] An image acquisition unit 310 for acquiring an original image;

[0110] An image processing unit 320 for preprocessing the original image to obtain a first image;

[0111] A feature point acquisition unit 330 for acquiring a first feature point set from the first image according to the Hessian matrix;

[0112] The feature point acquisition unit 330 is further configured to perform non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set;

[0113] The feature point acquisition unit 330 is further configured to enhance the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set.

[0114] In the embodiments of the present invention, the image processing unit 320 is further configured to:

[0115] Detect whether the original image is legal;

[0116] If the original image is illegal, output an empty set;

[0117] If the original image is legal, filter the original image to obtain the first image.

[0118] In an embodiment of the present invention, each first feature point in the first feature point set includes: a first key point and a corresponding BRIEF descriptor;

[0119] The feature point obtaining unit 330 is further configured to:

[0120] Obtain a set of first key points from the first image;

[0121] Extract the BRIEF descriptor of each of the first key points from the first image and the set of first key points through the BRIEF algorithm.

[0122] In an embodiment of the present invention, the key points include attributes, and the feature point obtaining unit 330 is further configured to:

[0123] Obtain a set of original feature points from the first image, where the set of original feature points includes a plurality of original feature points;

[0124] Obtain the Hessian matrix of each of the original feature points;

[0125] Obtain the characteristic response values of the Hessian matrix of the original feature points at multiple scale levels;

[0126] Use the original feature points corresponding to the characteristic response values greater than or equal to a preset threshold as first candidate feature points;

[0127] According to the corner enhancement algorithm, screen the first candidate feature points, and use the set of the screened first candidate feature points as a second candidate feature point set;

[0128] Obtain the attributes of each second candidate feature point in the second candidate feature point set, and use the set of second candidate feature points carrying attributes as the set of first key points.

[0129] In an embodiment of the present invention, the attribute includes a key point direction, and the feature point obtaining unit 330 is further configured to:

[0130] Obtain the centroid of a preset neighborhood of the second candidate feature points;

[0131] Use the direction from the second candidate feature points to the centroid as the key point direction.

[0132] In an embodiment of the present invention, the feature point obtaining unit 330 is further configured to:

[0133] Perform non-maximum suppression screening on one of the first feature points in the first feature point set to obtain a second feature point;

[0134] Among them, all the second feature points constitute the second feature point set.

[0135] In an embodiment of the present invention, the feature point acquisition unit 330 is further configured to:

[0136] Obtain a set of all first feature points within a preset range around the first feature point as a set of candidate feature points;

[0137] Obtain the feature response value of each first feature point in the set of candidate feature points;

[0138] Take the first feature point with the largest feature response value as the second feature point.

[0139] In an embodiment of the present invention, the feature point acquisition unit 330 is further configured to:

[0140] Perform denoising on the second feature point set to obtain a denoised point set;

[0141] On multiple scale levels, obtain a multi-scale point set from the denoised point set;

[0142] According to the corner enhancement algorithm, screen the multi-scale points in the multi-scale point set to obtain a corner point set;

[0143] After rotating and normalizing the key point directions of each corner point in the corner point set, obtain the enhanced feature point set.

[0144] The embodiment of the present invention can enhance feature points, especially improve the recognition degree of feature points in an application scenario with light changes, and avoid false matching.

[0145] The embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following method is implemented: obtain an original image; preprocess the original image to obtain a first image; according to the Hessian matrix, obtain a first feature point set from the first image; according to the MASK grid, perform non-maximum suppression screening on the first feature point set to obtain a second feature point set; according to the Feature Booster algorithm, enhance the second feature point set to obtain an enhanced feature point set.

[0146] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following method is implemented: obtaining an original image; preprocessing the original image to obtain a first image; obtaining a first set of feature points from the first image according to the Hessian matrix; performing non-maximum suppression screening on the first set of feature points according to a MASK grid to obtain a second set of feature points; enhancing the second set of feature points according to the FeatureBooster algorithm to obtain an enhanced set of feature points.

[0147] The above image feature point enhancement method realizes the beneficial effect of being able to solve the technical problems proposed in the background art.

[0148] Figure 2 It is a schematic flowchart of an image feature point enhancement method in an embodiment. It should be understood that although Figure 2 the steps in the flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, Figure 2 at least a part of the steps in

[0149] Figure 4 may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps. Figure 1 in Figure 4 shows the internal structure diagram of a computer device in an embodiment. The computer device may specifically be

[0150] the server 120 inFigure 4 The structure shown is only a block diagram of some of the structures related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0151] Those of ordinary skill in the art can understand that all or part of the processes of the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided by the present invention can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0152] It should be noted that in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0153] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather will be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for enhancing image feature points, characterized in that: The method comprises: Get the original image; Preprocessing the original image to obtain a first image; Obtaining a first feature point set from the first image according to the Hessian matrix; According to the MASK grid, non-maximum suppression screening is performed on the first feature point set to obtain a second feature point set; According to the Feature Booster algorithm, the second feature point set is enhanced to obtain an enhanced feature point set.

2. The method according to claim 1, characterized in that The preprocessing of the original image to obtain the first image includes: Detecting whether the original image is legal; If the original image is illegal, an empty set is output; If the original image is legal, the original image is filtered to obtain the first image.

3. The method according to claim 1, characterized in that Each first feature point in the first feature point set includes: a first key point and a corresponding BRIEF descriptor; The step of obtaining a first feature point set from the first image according to the Hessian matrix includes: Acquire a first key point set from the first image; A BRIEF descriptor of each of the first key points is extracted from the first image and the first key point set by using a BRIEF algorithm.

4. The method according to claim 3, characterized in that: The key point includes an attribute, and acquiring a first key point set from the first image includes: Acquire an original feature point set from the first image, wherein the original feature point set includes a plurality of original feature points; Obtain the Hessian matrix of each of the original feature points; At multiple scale levels, obtaining characteristic response values ​​of the Hessian matrix of the original feature points; The original feature point whose corresponding feature response value is greater than or equal to a preset threshold is used as a first candidate feature point; The first candidate feature points are screened according to a corner point enhancement algorithm, and a set of the screened first candidate feature points is used as a second candidate feature point set; The attribute of each second candidate feature point in the second candidate feature point set is obtained, and the second candidate feature point set carrying the attribute is used as the first key point set.

5. The method according to claim 4, characterized in that The attribute includes a key point direction, and obtaining the attribute of each second candidate feature point in the second candidate feature point set includes: Obtaining the centroid of a preset neighborhood of the second candidate feature point; The direction from the second candidate feature point to the centroid is used as the key point direction.

6. The method according to claim 1, characterized in that The step of performing non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set includes: Performing non-maximum suppression screening on one of the first feature points in the first feature point set to obtain a second feature point; Among them, all the second feature points constitute the second feature point set.

7. The method according to claim 6, characterized in that The performing non-maximum suppression screening on one of the first feature points in the first feature point set to obtain a second feature point includes: Acquire a set of all first feature points within a preset range around the first feature point as a candidate feature point set; Obtaining a feature response value of each of the first feature points in the set of candidate feature points; The first feature point with the largest feature response value is used as the second feature point.

8. The method according to claim 1, characterized in that The step of enhancing the second feature point set according to the Feature Booster algorithm to obtain an enhanced feature point set includes: De-noising the second feature point set to obtain a de-noised point set; At multiple scale levels, obtaining a multi-scale point set from the denoised point set; According to a corner point enhancement algorithm, the multi-scaled points in the multi-scaled point set are screened to obtain a corner point set; After the key point direction of each corner point in the corner point set is rotationally normalized, the enhanced feature point set is obtained.

9. An image feature point enhancement device, characterized in that: The device comprises: An image acquisition unit, used for acquiring an original image; An image processing unit, used for preprocessing the original image to obtain a first image; A feature point acquisition unit, configured to acquire a first feature point set from the first image according to a Hessian matrix; The feature point acquisition unit is further used to perform non-maximum suppression screening on the first feature point set according to the MASK grid to obtain a second feature point set; The feature point acquisition unit is further used to enhance the second feature point set according to a Feature Booster algorithm to obtain an enhanced feature point set.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images