Pairing analysis method, device and equipment for kernel objects combined in pairs and medium

Abstract

The invention relates to a pairing analysis method, device and equipment for kernel objects combined in pairs and a medium, and the method comprises the steps: extracting first contour data and second contour data based on first image data corresponding to a current paired object and second image data corresponding to a target paired object; based on the first contour data and the second contour data, obtaining an average contour similarity between the current paired object and the target paired object; based on the first texture data corresponding to the current paired object and the second texture data corresponding to the target paired object, obtaining the texture similarity between the current paired object and the target paired object; and on the basis of the average contour similarity and the texture similarity, comprehensive similarity information and a pairing analysis result between the current pairing object and the target pairing object are acquired. According to the method, the similarity of the contours is comprehensively evaluated in multiple directions, and then comprehensive analysis is performed based on the line similarity, so that the similarity of the paired objects can be comprehensively measured, and the pairing accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular, to a pairing analysis method, device, equipment and medium for paired nuclear objects. Background Art

[0002] Paired nuclear objects are often important carriers of traditional collections and cultures, and their pairing quality directly affects the collection value and collection experience. Paired nuclear objects can include walnuts. In the collection market, the pairing of paired nuclear objects needs to comprehensively consider multi-dimensional features such as shape, texture, size, and cortex. Traditional methods rely on manual experience and achieve pairing through visual comparison or measurement with measuring tools. However, manual pairing highly depends on the experience accumulation of appraisers, and there are subjective differences in the judgment criteria of "similarity" among different appraisers, resulting in poor consistency of pairing results, time-consuming and laborious, difficult to meet the needs of large-scale transactions, and having technical defects of strong subjectivity and low efficiency.

[0003] Therefore, there is an urgent need for an automated pairing analysis solution for nuclear objects that can integrate multi-dimensional features and quantify pairing similarity to solve the efficiency and accuracy bottlenecks of traditional manual pairing and promote the standardization and intelligent upgrade of nuclear object pairing. Summary of the Invention

[0004] In order to improve the efficiency and accuracy of nuclear object pairing and promote the standardization and intelligent upgrade of nuclear object pairing, the present application provides a pairing analysis method, device, equipment and medium for paired nuclear objects.

[0005] In a first aspect, the present application provides a pairing analysis method for paired nuclear objects, including: For each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations; For each target paired object at the current orientation, based on the first contour data corresponding to the current orientation and the second contour data corresponding to the current orientation, perform center point alignment processing, and calculate the single contour similarity between the current paired object and the target paired object in the current orientation; For each target paired object, based on the single contour similarities corresponding to different orientations respectively, obtain the average contour similarity between the current paired object and the target paired object; Based on the first image data, extract the first texture data corresponding to the current paired object. For each target paired object, based on the second image data, extract the second texture data corresponding to the target paired object. Based on the first texture data and the second texture data, obtain the texture similarity between the current paired object and the target paired object. For each target paired object, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object.

[0006] The beneficial effects of the present invention are as follows: By automatically extracting and analyzing the contour and texture data, the pairing result can be obtained quickly, saving time and manpower. The similarity of the contours is comprehensively evaluated in multiple aspects, and then comprehensive analysis is performed based on the texture similarity, which can comprehensively measure the similarity of the paired objects and improve the pairing accuracy. Through image processing and feature extraction technologies, the noise and interference are effectively reduced, making the pairing result more stable and reliable. At the same time, it can be applied to paired objects of different varieties and shapes, thus having wide adaptability and application prospects.

[0007] On the basis of the above technical solutions, the present invention can also be improved as follows.

[0008] Further, before for each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object in different orientations and the second contour data corresponding to the target paired object in different orientations, it further includes: Based on the first image data, obtain the first screening information corresponding to the current paired object, and the first screening information includes the first size information and the first color information; Based on the second image data corresponding to each of the multiple candidate paired objects, obtain the second screening information corresponding to each of the candidate paired objects, and each second screening information includes the second size information and the second color information; Based on the first screening information and the multiple second screening information, screen the multiple candidate paired objects, and use the screened candidate paired objects as the target paired objects respectively.

[0009] The beneficial effect of adopting the above further solution is: By pre-excluding the candidate paired objects that are obviously mismatched, resources can be concentrated to perform detailed analysis on the target paired objects that are more likely to be successfully paired, improving the efficiency of the entire pairing process and effectively reducing the computational amount and time cost of subsequent precise pairing analysis.

[0010] Further, before obtaining the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object for each of the target paired objects based on the average contour similarity and the texture similarity, the following steps are also included: Based on the third image data corresponding to the current paired object, obtain the first area value and the first gray value corresponding to the first illuminated area, and calculate the first surface smoothness degree corresponding to the current paired object based on the first area value and the first gray value; the third image data is the image data collected under the illumination of an auxiliary light source at a preset angle. For each of the target paired objects, based on the fourth image data corresponding to the target paired object, obtain the second area value and the second gray value corresponding to the second illuminated area, and calculate the second surface smoothness degree corresponding to the target paired object based on the second area value and the second gray value; the fourth image data is the image data collected under the illumination of the auxiliary light source. For each of the target paired objects, calculate the gloss similarity between the current paired object and the target paired object based on the first surface smoothness degree and the second surface smoothness degree.

[0011] The beneficial effect of adopting the above further solution is: it improves the convenience and accuracy of obtaining the gloss similarity, and is convenient for subsequent pairing of the current paired object in combination with the gloss similarity.

[0012] Further, for each of the target paired objects, obtaining the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object based on the average contour similarity and the texture similarity includes: Based on the target requirements and attribute information corresponding to the current paired object, obtain the first weight information corresponding to the average contour similarity, the texture similarity, and the gloss similarity respectively, where the target requirements are parameters representing the user's consistency requirements for paired objects, and the attribute information is parameters representing the physical characteristics of paired objects; Based on each of the first weight information, perform weighted calculation on the average contour similarity, the texture similarity, and the gloss similarity to obtain the comprehensive similarity information, and obtain the pairing analysis result based on the comprehensive similarity information, where the pairing analysis result includes pairing success and pairing failure.

[0013] The beneficial effect of adopting the above further solution is: the calculation of the comprehensive similarity information can comprehensively and accurately reflect the similarity of paired objects in multiple key dimensions. By comparing with the similarity threshold, the pairing analysis result can be quickly and clearly given, providing a clear and reliable pairing judgment basis for users.

[0014] Further, before, for each target paired object, extracting the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, further includes: Based on the fifth image data corresponding to the current paired object, obtaining first current angle information of the current paired object in a preset coordinate system; For each of the target paired objects, based on the sixth image data corresponding to the target paired object, obtaining second current angle information of the target paired object in a preset coordinate system; If the first current angle information conforms to preset first target angle information, then execute the step of collecting the first image data; if the second current angle information conforms to preset second target angle information, then execute the step of collecting the second image data.

[0015] The beneficial effects of adopting the above further solution are: it can ensure that when extracting contour data, both the current paired object and the target paired object are in the preset optimal angle positions. It reduces the possibility of errors in contour extraction caused by angle deviations of the paired objects, and improves the reliability and accuracy of the entire pairing analysis process.

[0016] Further, for each of the target paired objects, the obtaining the texture similarity between the current paired object and the target paired object based on the first texture data and the second texture data includes: For each of the target paired objects, performing an overlapping process on the first texture data and the second texture data, obtaining the intersection texture area and the union texture area between the first texture data and the second texture data, and calculating a first sub-texture similarity between the current paired object and the target paired object based on the intersection texture area and the union texture area; For each of the target paired objects, based on a trained texture change prediction model, the first texture data, and the second texture data, predicting third texture data corresponding to the current paired object and fourth texture data corresponding to the target paired object, where the texture data includes color features, shape features, and distribution features; For each of the target paired objects, calculating a second sub-texture similarity between the current paired object and the target paired object based on the third texture data and the fourth texture data; For each of the target paired objects, obtaining the texture similarity based on the first sub-texture similarity and the second sub-texture similarity.

[0017] The beneficial effects of adopting the above further solution are as follows: By comprehensively considering the similarity of the first sub-pattern and the similarity of the second sub-pattern, the final pattern similarity is obtained. By integrating the actual and predicted similarity data, a more comprehensive and accurate pattern similarity can be obtained, providing a more reliable basis for the pairing decision.

[0018] Further, for each of the target paired objects, based on the third pattern data and the fourth pattern data, calculating the second sub-pattern similarity between the current paired object and the target paired object includes: Based on the first pattern data, the second pattern data, the third pattern data, and the fourth pattern data, determining the respective change degrees corresponding to the color feature, the shape feature, and the distribution feature in the pattern data; Based on each of the change degrees and a preset weight adjustment rule, respectively determining the second weight information corresponding to the color feature, the shape feature, and the distribution feature; the weight adjustment rule includes that within the weight range corresponding to each feature, the greater the change degree of the feature, the higher the corresponding second weight information; Based on the pattern change prediction model, obtaining the prediction uncertainties for the color feature, the shape feature, and the distribution feature; Based on each of the prediction uncertainties, respectively adjusting the second weight information corresponding to the color feature, the shape feature, and the distribution feature to obtain the third weight information, and calculating the second sub-pattern similarity based on the third weight information.

[0019] The beneficial effects of adopting the above further solution are as follows: By considering the change degrees of each pattern feature during the process of playing with the object, the weights are dynamically adjusted, enabling the calculation of the second sub-pattern similarity to better reflect the actual pattern changes. The influence of prediction errors on the calculation of the second sub-pattern similarity is reduced, and the reliability of the calculation of the second sub-pattern similarity is improved.

[0020] In a second aspect, the present application provides a pairing analysis device for a pair of combined nuclear objects, including: An extraction contour module, configured to, for each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations; An obtaining single contour similarity module, configured to, for each target paired object at the current orientation, based on the first contour data corresponding to the current orientation and the second contour data corresponding to the current orientation, perform a center point alignment process, and calculate the single contour similarity between the current paired object and the target paired object in the current orientation; An average contour similarity obtaining module, configured to obtain, for each of the target paired objects, an average contour similarity between the current paired object and the target paired object based on the single contour similarities corresponding to different orientations; A texture similarity obtaining module, configured to extract first texture data corresponding to the current paired object based on the first image data, extract second texture data corresponding to each of the target paired objects based on the second image data, and obtain a texture similarity between the current paired object and each of the target paired objects based on the first texture data and the second texture data; A comprehensive analysis module, configured to obtain, for each of the target paired objects, comprehensive similarity information and a pairing analysis result between the current paired object and the target paired object based on the average contour similarity and the texture similarity.

[0021] In a third aspect, the present application provides an electronic device, including a processor and a memory, where the processor is coupled to the memory; The processor is configured to execute a computer program stored in the memory, so that the electronic device executes the method according to any one of the first aspect.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium, including a computer program or instructions, where when the computer program or instructions run on a computer, the computer is caused to execute the method according to any one of the first aspect. Description of the Drawings

[0023] Figure 1 It is a schematic flowchart of a pairing analysis method for nuclear objects in paired combination according to an embodiment of the present application; Figure 2 It is a schematic diagram showing the approximate areas of paired objects according to an embodiment of the present application; Figure 3 It is a schematic diagram of the contour data of paired objects according to an embodiment of the present application; Figure 4 It is a schematic diagram of the texture data of paired objects according to an embodiment of the present application; Figure 5 It is a structural block diagram of a pairing analysis device for nuclear objects in paired combination according to an embodiment of the present application; Figure 6 It is a structural block diagram of an electronic device according to an embodiment of the present application. Detailed Embodiments

[0024] The principles and features of the present invention are described below. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0025] The present application will be further described in detail below with reference to the accompanying drawings.

[0026] An embodiment of the present application provides a pairing analysis method for paired nuclear objects. This method can be executed by a device, which can be a server or a terminal device. The server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smart phone, a tablet computer, a desktop computer, etc., but is not limited thereto.

[0027] As Figure 1 shown, a pairing analysis method for paired nuclear objects, with an electronic device as the execution subject, the main process of the method is described as follows (Steps S101 - S105): Step S101: For each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object in different orientations and the second contour data corresponding to the target paired object in different orientations.

[0028] In this embodiment, paired nuclear objects refer to drupe - like objects that need to be paired in groups of two. Drupe - like objects can include walnuts for cultural relics, olive pits, etc. In certain specific usage scenarios or cultural backgrounds, paired nuclear objects often need to find paired individuals with similar shapes, textures, colors, etc.

[0029] The paired object is a paired nuclear object, the current paired object is the nuclear object that needs to be paired currently, and the target paired object is the nuclear object to be selected corresponding to the current paired object. The target paired object corresponding to the current paired object can be one or multiple.

[0030] The first image data refers to a set of images of the current paired object taken from multiple different orientations, and the second image data refers to a set of images of the target paired object taken from multiple different orientations. When the paired object is a walnut for cultural relics, the first image data includes images of the current walnut for cultural relics taken from different angles, and the second image data includes images of the target walnut for cultural relics taken from different angles. Through image data from multiple orientations, the contour of the paired object can be comprehensively reflected, which can provide rich information for the subsequent extraction of contour data. In this embodiment, when the paired object is a walnut for cultural relics, since the walnut for cultural relics can be divided into six faces, the first image data and the second image data can respectively include images of six orientations.

[0031] Based on image processing technology, the contour of the paired object can be accurately identified from the captured first image data and second image data.

[0032] Specifically, first, since there may be a turntable and other background noise in the image of the paired object, it is necessary to firstly compare the first image data and the second image data, such as Figure 2 As shown, the images in the same orientation are preprocessed, and the preprocessing may include graying, Gaussian blurring and other operations to obtain the approximate area of ​​the paired objects. When shooting, the paired objects are placed on a turntable, and the approximate area of ​​the objects may be an area composed of the object area and the turntable. Next, the edge detection algorithm is used to detect the edges of the two paired objects in the same orientation, and the polygonal approximation method is used to fit the contour shape of the paired objects, so that the contour data of the two paired objects in different orientations can be obtained.

[0033] Step S102: For each target paired object in the current position, center point alignment is performed based on the first contour data corresponding to the current position and the second contour data corresponding to the current position, and a single contour similarity between the current paired object and the target paired object in the current position is calculated.

[0034] In this embodiment, based on the preset first formula, the similarity of the single contours of two paired objects in the same specific orientation can be calculated. Specifically, Figure 3 As shown, the first contour data and the second contour data are aligned with each other, and then the distance between each point in the two contour data is calculated to find the maximum difference value. It is easy to understand that the center point alignment process is to correct the position of the contour data of two paired objects so that their center points coincide with each other. The center point can be the geometric center of the contour data.

[0035] Maximum difference It is a key indicator to measure the difference between two contours, and refers to the maximum value of the distance between the corresponding points on the contours of the current paired object and the target paired object. Maximum difference value The larger it is, the greater the difference between the two contour shapes.

[0036] Based on the first formula, the maximum difference value when obtaining the current position The corresponding single contour similarity. The first formula can be expressed as:

[0037] in, Indicates the similarity of a single contour corresponding to the current orientation; Indicates the maximum difference between two contour data obtained from historical tests; represents the contour weight coefficient corresponding to the current orientation, that is, the weight coefficient corresponding to the specific face of the paired object displayed at this orientation. In this embodiment, It can be the maximum difference value obtained after testing thousands of contour data.

[0038] Step S103: For each of the target paired objects, based on the individual contour similarities corresponding to different orientations, obtain the average contour similarity between the current paired object and the target paired object.

[0039] In this embodiment, based on a preset second formula, the numerical values of the individual contour similarities in multiple different orientations are averaged to obtain the average contour similarity. Thus, it can comprehensively reflect the similarity of the shapes of two matching objects at different angles, reducing the possibility of affecting the overall judgment due to the contingency of a single angle.

[0040] Specifically, by accumulating the individual contour similarities in each orientation, the total accumulated score corresponding to all orientations can be obtained , and based on the second formula, obtain the current total accumulated score and the corresponding average contour similarity. The second formula can be expressed as:

[0041] where represents the average contour similarity; represents the number of contours of each paired object. Since for each orientation, there is a corresponding number of contours of this paired object, the number of contours is equal to the number of orientations of the image capture. Exemplarily, when the paired object is a walnuts for cultural relics, since the walnuts for cultural relics can be divided into six faces, and each face corresponds to a contour, the number of contours can be set to 6.

[0042] Through the average contour similarity , the contour similarity between two paired objects in all key orientations can be comprehensively reflected.

[0043] Step S104: Based on the first image data, extract the first texture data corresponding to the current paired object. For each of the target paired objects, based on the second image data, extract the second texture data corresponding to the target paired object. Based on the first texture data and the second texture data, obtain the texture similarity between the current paired object and the target paired object.

[0044] Such as Figure 4As shown, during the extraction of the texture data, it is necessary to preprocess the collected color image first. Specifically, the image is converted from the original color space to the standard RGB color space to ensure the accuracy and consistency of the color information. After the conversion, image enhancement techniques can be used to process the belly surface area of the paired objects, enhancing the texture features in the concave areas of the belly surface area to make them more obvious. Image enhancement methods can include histogram equalization, contrast adjustment, etc. After the enhancement process, the concave areas in the belly surface area form a more distinct contrast with the surrounding areas. Then, the image is binarized, and the pixel points in the image are divided into two categories, foreground and background, according to the set threshold. The foreground pixels represent the concave areas in the belly surface area, while the background pixels represent the other parts. Through this series of processes, the texture area of the belly surface area and the texture data corresponding to the texture area can be successfully extracted, providing a clear and accurate data basis for subsequent texture similarity analysis.

[0045] It is easy to understand that the belly surface area refers to the abdominal position of the nuclear object, which is a region that bulges outward as a whole. When the paired object is a walnuts for fun, the main veins of the walnuts for fun can divide the main body of the walnut into multiple belly surface areas. The main vein refers to the main vein that extends from the top to the bottom of the walnuts for fun. Exemplarily, the main vein of the walnuts for fun can be one, thus including two belly surface areas.

[0046] Based on the first texture data and the second texture data, the texture similarity between the current paired object and the target paired object can be obtained.

[0047] As an alternative implementation of this embodiment, the first texture data and the second texture data can be obtained separately according to the orientations of different belly surface areas, so as to calculate the individual similarities corresponding to multiple different belly surface areas respectively. Similar to the above logic of calculating the average contour similarity, the multiple individual similarities can be accumulated and then averaged to obtain the texture similarity between the current paired object and the target paired object.

[0048] For example, when the paired object is a walnuts for fun, the texture data collection and individual similarity calculation can be performed on multiple belly surface areas of the walnuts for fun respectively to ensure that the textures at each key position of the two walnuts for fun can be accurately compared and evaluated.

[0049] As another alternative implementation of this embodiment, if it is possible to determine that the texture data of multiple belly surface areas in the same paired object are relatively similar according to the material information of the paired object, then the key belly surface area among the multiple belly surface areas can be selected, and the first texture data corresponding to the key belly surface area of the current paired object and the second texture data corresponding to the key belly surface area of the target paired object are obtained. Based on the first texture data and the second texture data, the texture similarity between the current paired object and the target paired object is obtained.

[0050] Step S105: For each of the target paired objects, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object.

[0051] Calculate the comprehensive similarity information of the current walnut and the target walnut according to the average contour similarity and the texture similarity. And the pairing analysis result can be obtained by judging whether the pairing is successful according to the set similarity threshold.

[0052] In this embodiment, the pairing analysis method can achieve efficient and accurate pairing of paired objects, meeting the user's requirements for high-precision and high-consistency pairing of paired objects. Specifically, by automatically extracting and analyzing the contour and texture data, the pairing result can be quickly obtained, saving time and manpower. The similarity of the contour is comprehensively evaluated in multiple directions, and then comprehensive analysis is performed based on the texture similarity, which can comprehensively measure the similarity of paired objects and improve the pairing accuracy. Through image processing and feature extraction technologies, the noise and interference are effectively reduced, making the pairing result more stable and reliable. At the same time, it can be applied to paired objects of different varieties and shapes, thus having wide adaptability and application prospects.

[0053] In this embodiment, before the step S101, the following processing may further be included: based on the first image data, obtain the first screening information corresponding to the current paired object, where the first screening information includes the first size information and the first color information; based on the second image data respectively corresponding to multiple candidate paired objects, obtain the second screening information respectively corresponding to each of the candidate paired objects, and each of the second screening information includes the second size information and the second color information; based on the first screening information and the multiple second screening information, screen the multiple candidate paired objects, and use the screened candidate paired objects as the target paired objects respectively.

[0054] By analyzing image processing technologies such as size measurement algorithms and color histograms, the first screening information corresponding thereto can be obtained based on the first image data of the current paired object, and the second screening information corresponding thereto can be obtained based on the second image data of the candidate paired object. The size information in the screening information may respectively include the height, width, thickness, etc. of the current paired object, and the color information may respectively include the RGB three-channel color values, etc. When the paired object is a walnuts for cultural relics collection, the size information may further include the width of the rib and the width of the belly. It is easy to understand that the width of the rib refers to the width of the main rib in the walnuts for cultural relics collection, and the width of the belly refers to the width of the belly area in the walnuts for cultural relics collection.

[0055] Use the first screening information and the second screening information of all candidate paired objects to preliminarily screen all candidate walnuts. Specifically, by setting similarity thresholds for size and color, candidate paired objects that differ too much from the first size information and the first color information can be excluded, and the remaining candidate paired objects are regarded as target paired objects.

[0056] By pre-excluding candidate paired objects that are obviously mismatched, resources can be concentrated to conduct a detailed analysis of target paired objects that are more likely to be successfully paired, improving the efficiency of the entire pairing process and effectively reducing the computational amount and time cost of subsequent precise pairing analysis.

[0057] In this embodiment, before the step S105, the following processing may further be included: Based on the third image data corresponding to the current paired object, obtain the first area value and the first gray value corresponding to the first illuminated area, and calculate the first surface smoothness corresponding to the current paired object based on the first area value and the first gray value; the third image data is the image data collected based on the illumination of an auxiliary light source at a preset angle. For each of the target paired objects, based on the fourth image data corresponding to the target paired object, obtain the second area value and the second gray value corresponding to the second illuminated area, and calculate the second surface smoothness corresponding to the target paired object based on the second area value and the second gray value; the fourth image data is the image data collected based on the illumination of the auxiliary light source. For each of the target paired objects, calculate the gloss similarity between the current paired object and the target paired object based on the first surface smoothness and the second surface smoothness.

[0058] In this embodiment, the mirror reflection principle can be used, and an auxiliary light source is used to irradiate the surface of the paired object at a preset angle, and then the illuminated third image data or fourth image data is collected by a camera. The auxiliary light source can be a 50W strip light source, and the preset angle can be that the auxiliary light source forms a 45-degree angle with the camera. The third image data or the fourth image data is preprocessed by converting it into a grayscale image and Gaussian blurring to obtain the area value and the gray value of the illuminated area. Based on a preset third formula, by synthesizing the area value and the gray value of the illuminated area, the surface smoothness of the paired object can be obtained.

[0059] The third formula can be expressed as:

[0060] Wherein, represents the surface smoothness; represents the area value of the illuminated area; represents the total area of the corresponding paired objects; represents the median of the gray values of the illuminated area; The result is the sum of the scores of the illuminated part and the gray value score. 50% represents the weight of the corresponding illuminated part score or gray value score. The weights of the two can be the same, that is, both are 50%, or they can be other inconsistent weight values; HD / 255*100 represents the conversion method of converting the gray value from 0 - 255 to 0 - 100 points.

[0061] Based on the preset fourth formula, the first surface smoothness, and the second surface smoothness, the gloss similarity between the current paired object and the target paired object can be obtained. The fourth formula can be expressed as:

[0062] wherein, represents the gloss similarity between two paired objects; represents the first surface smoothness corresponding to the current paired object; represents the second surface smoothness corresponding to the target paired object. The difference between the first surface smoothness and the second surface smoothness represents the gloss similarity. The smaller the difference, the higher the similarity between the two.

[0063] In this embodiment, the step S105 may specifically include the following processing: Based on the target requirements and attribute information corresponding to the current paired object, obtain the first weight information corresponding to each of the average contour similarity, the texture similarity, and the gloss similarity. The target requirement is a parameter representing the user's consistency requirement for the paired object, and the attribute information is a parameter representing the physical characteristics of the paired object; Based on each of the first weight information, perform weighted calculation on the average contour similarity, the texture similarity, and the gloss similarity to obtain the comprehensive similarity information, and based on the comprehensive similarity information, obtain the pairing analysis result, where the pairing analysis result includes pairing success and pairing failure.

[0064] In this embodiment, the target requirement refers to the user's requirement for the consistency of the paired object in terms of appearance, etc. For example, the user may pay more attention to the consistency of the texture or the matching degree of the gloss. The attribute information refers to the physical characteristics of the paired object itself, such as the material. According to the target requirements and attribute information, corresponding first weight information can be assigned to each similarity index to reflect their importance in the comprehensive evaluation.

[0065] By assigning weights to different similarity indexes, the comprehensive similarity information can be made more in line with the actual needs of the user and the characteristics of the paired object, improving the pertinence and satisfaction of the pairing result.

[0066] Based on a preset fifth formula, the average contour similarity, texture similarity, and gloss similarity can be calculated with weights. According to the set similarity threshold, it is determined whether the pairing is successful. If the comprehensive similarity information reaches or exceeds the threshold, the pairing is successful; otherwise, the pairing fails.

[0067] The fifth formula can be expressed as:

[0068] Wherein, represents the comprehensive similarity information; represents the basic score assigned to each target paired object. For example, can be 50; represents the average contour similarity; represents the first weight information corresponding to the average contour similarity; represents the texture similarity; represents the first weight information corresponding to the texture similarity; represents the gloss similarity; represents the first weight information corresponding to the gloss similarity.

[0069] In this embodiment, the calculation of the comprehensive similarity information can comprehensively and accurately reflect the similarity of the paired objects in multiple key dimensions. By comparing with the similarity threshold, the pairing analysis result can be given quickly and clearly, providing a clear and reliable basis for pairing judgment for users.

[0070] In this embodiment, before the step S101, the following processing may further be included: based on the fifth image data corresponding to the current paired object, obtaining the first current angle information of the current paired object in a preset coordinate system; for each of the target paired objects, based on the sixth image data corresponding to the target paired object, obtaining the second current angle information of the target paired object in the preset coordinate system; if the first current angle information conforms to the preset first target angle information, then execute the step of collecting the first image data; if the second current angle information conforms to the preset second target angle information, then execute the step of collecting the second image data.

[0071] Based on the fifth image data of the current paired object, through a preset angle measurement algorithm, the first current angle information of the current paired object is calculated. Based on the sixth image data of the target paired object, through the preset angle measurement algorithm, the second current angle information of the current paired object is calculated. The angle measurement algorithm can be a deep learning model of the angle detection type. By using a labeling tool to label the pictures of the paired objects to generate a data set, this deep learning model can be obtained through training.

[0072] When the paired object is a walnut for cultural relics collection, the fifth image data may be an image of the top of the walnut for cultural relics collection. Through the trained deep learning model, the angle of the main veins of the walnut for cultural relics collection in the fifth image data or the sixth image data is identified, and the angle of the main veins is used as the corresponding current angle information. In this embodiment, the angle formed by the connection line between the two points with the farthest distance in the main veins in a preset coordinate system can be used as the angle of the main veins.

[0073] Through the judgment and acquisition steps of the angle information, it can be ensured that when extracting the contour data, both the current paired object and the target paired object are in the preset optimal angle positions. The possibility of errors in contour extraction caused by the angle deviation of the paired object is reduced, and the reliability and accuracy of the entire pairing analysis process are improved.

[0074] As an optional implementation manner in this embodiment, for each of the target paired objects, obtaining the texture similarity between the current paired object and the target paired object based on the first texture data and the second texture data may include the following processing: for each of the target paired objects, overlapping the first texture data and the second texture data, obtaining the intersection texture area and the union texture area between the first texture data and the second texture data, and calculating the first sub-texture similarity between the current paired object and the target paired object based on the intersection texture area and the union texture area, and using the first sub-texture similarity as the texture similarity.

[0075] After extracting the first texture data and the second texture data, it is necessary to perform an alignment operation on the texture data of these two paired objects. The alignment operation may be to make the second texture data of the second paired object coincide with the first texture data of the first paired object in position through translation transformation according to the coordinates of the center point of the first paired object.

[0076] After alignment, the calculation methods of intersection and union can be used to quantify the similarity degree of the textures. Specifically, the intersection texture area and the union texture area of the two texture data can be calculated. The intersection texture area represents the area of the overlapping part of the two texture data, that is, the area where their textures are the same; while the union texture area represents the total area of the two texture data.

[0077] By using a preset sixth formula to calculate the ratio of the intersection texture area to the union texture area and multiplying this ratio by 100%, a percentage representation of the texture similarity between two paired objects can be obtained. Through this percentage representation of texture similarity, the degree of consistency between the two texture data can be intuitively reflected. When the paired objects are walnuts used as cultural relics, if the currently recognized first texture data and second texture data are exactly the same, the intersection texture area and the union texture area will be equal, and the texture similarity at this time is 100%, that is, they are completely similar.

[0078] The sixth formula can be expressed as:

[0079] Among them, represents the first sub-texture similarity; represents the intersection texture area; represents the union texture area. The first sub-texture similarity is the texture similarity .

[0080] As another alternative implementation in this embodiment, for each of the target paired objects, obtaining the texture similarity between the current paired object and the target paired object based on the first texture data and the second texture data may include the following processes: For each of the target paired objects, perform an overlapping process on the first texture data and the second texture data to obtain the intersection texture area and the union texture area between the first texture data and the second texture data, and calculate the first sub-texture similarity between the current paired object and the target paired object based on the intersection texture area and the union texture area; For each of the target paired objects, based on the trained texture change prediction model, the first texture data, and the second texture data, predict the third texture data corresponding to the current paired object and the fourth texture data corresponding to the target paired object, where the texture data includes color features, shape features, and distribution features; For each of the target paired objects, calculate the second sub-texture similarity between the current paired object and the target paired object based on the third texture data and the fourth texture data; For each of the target paired objects, obtain the texture similarity based on the first sub-texture similarity and the second sub-texture similarity.

[0081] In this alternative embodiment, the calculation process of the first sub-pattern similarity is the same as that in the above alternative embodiment, and will not be elaborated here. As the paired objects are played with, the corresponding texture colors, texture shapes, texture distributions, etc. may change. Therefore, in this alternative embodiment, a trained pattern change prediction model is used to predict the third pattern data corresponding to the current paired object and the fourth pattern data corresponding to the target paired object based on the first pattern data of the current paired object and the second pattern data of the target paired object. The pattern change prediction model considers the change trends in multiple dimensions such as color features, shape features, and distribution features. Through the pattern change prediction model, the change situation of the pattern data of the paired objects at different stages can be understood in advance, providing data support for evaluating the long-term similarity of the patterns.

[0082] The pattern change prediction model is trained based on a large amount of pattern data. By analyzing the changes in the pattern characteristics of the paired objects at different playing stages, including multiple dimensions such as color, shape, and distribution, a dynamic model of pattern change is established. The large amount of pattern data can be sourced from observations and records during the actual playing process, covering the pattern changes of paired objects of different varieties and different maturities at different playing times.

[0083] Based on the predicted third pattern data and fourth pattern data, the second sub-pattern similarity between the current paired object and the target paired object is obtained. By introducing the similarity calculation of the predicted data, the forward-looking and comprehensiveness of the evaluation are increased, realizing the prediction and evaluation of the pattern similarity in the future or under different conditions, which helps to identify the pairings of paired objects that can still maintain a high similarity under specific change conditions.

[0084] In this alternative embodiment, the final pattern similarity can be obtained by performing a weighted calculation on the first sub-pattern similarity and the second sub-pattern similarity. By integrating the actual and predicted similarity data, a more comprehensive and accurate pattern similarity can be obtained, providing a more reliable basis for pairing decisions.

[0085] In this alternative embodiment, for each of the target paired objects, calculating the second sub-pattern similarity between the current paired object and the target paired object based on the third pattern data and the fourth pattern data may include the following processing: Based on the first pattern data, the second pattern data, the third pattern data, and the fourth pattern data, determine the respective degrees of change of the color feature, shape feature, and distribution feature in the pattern data; Based on each of the degrees of change and a preset weight adjustment rule, respectively determine the second weight information corresponding to the color feature, the shape feature, and the distribution feature; the weight adjustment rule includes that within the weight range corresponding to each feature, the greater the degree of change of the feature, the higher the corresponding second weight information. Based on the texture change prediction model, obtain the prediction uncertainties of the color feature, the shape feature, and the distribution feature. Based on each of the prediction uncertainties, respectively adjust the second weight information corresponding to the color feature, the shape feature, and the distribution feature to obtain the third weight information, and based on the third weight information, calculate the second sub-texture similarity.

[0086] For each pair of paired objects, the predicted values of the degrees of change of each texture feature during the rubbing process can be obtained by using the third texture data and the fourth texture data. The degree of change is a parameter characterizing the difference between each feature in the predicted texture data and the original texture data. For example, it is predicted that the degree of change of the color feature of the current paired object is 80 (the larger the value, the greater the change), the degree of change of the shape feature is 50, and the degree of change of the distribution feature is 60; the degree of change of the color feature of the target paired object is 70, the degree of change of the shape feature is 40, and the degree of change of the distribution feature is 50.

[0087] In this alternative embodiment, the relative importance of each texture feature can be calculated according to the degree of change of each texture feature. Taking the color feature as an example, the sum of the degrees of change of the color features of the current paired object and the target paired object is 80 + 70 = 150. The relative importance of the color feature of the current paired object is 80 / 150 ≈ 0.533, and the relative importance of the color feature of the target paired object is 70 / 150 ≈ 0.467. Then, within the preset weight range, the specific weight value can be determined according to the relative importance. For example, if the weight range of the color feature is from 0.1 to 0.5, the weight of the color feature of the current paired object can be set to 0.5×0.533 + 0.1×(1 - 0.533) ≈ 0.343, and the weight of the color feature of the target paired object is set to 0.5×0.467 + 0.1×(1 - 0.467) ≈ 0.283. Similarly, similar calculations are performed on the shape feature and the distribution feature respectively to obtain the initial second weight information corresponding to each texture feature. The initial second weight information can also be normalized to obtain the final second weight information for each to ensure reasonable weight distribution and a total sum of 1.

[0088] For each texture feature, its prediction uncertainty is calculated by analyzing the output probability distribution of the texture change prediction model or the variance of the prediction results, etc. For example, when using a Bayesian neural network for texture change prediction, the probability distribution of the prediction results for each texture feature change can be obtained. Taking the prediction of color feature change as an example, the texture change prediction model outputs a probability of 0.6 that the color feature becomes dark brown after being played with, a probability of 0.3 for reddish-brown, and a probability of 0.1 for light brown. At this time, the entropy of the prediction result of this color feature is - (0.6×log0.6 + 0.3×log0.3 + 0.1×log0.1) ≈ 0.881. The larger the entropy value, the higher the uncertainty of the prediction result.

[0089] Set the adjustment strategy of the second weight information according to the prediction uncertainty. For example, for features with higher prediction uncertainty, their weights can be reduced by a certain proportion. An adjustment coefficient can be set. For example, when the prediction uncertainty exceeds a certain threshold, the reduction ratio of the weight is the adjustment coefficient multiplied by the excess part of the uncertainty. Assume that the set entropy threshold is 0.5 and the adjustment coefficient is 0.5. For the above color feature, its entropy value is 0.881, and the excess part over 0.5 is 0.381, then the weight reduction ratio is 0.5×0.381 ≈ 0.1905.

[0090] Adjust the second weight information of each texture feature according to the adjustment strategy of the second weight information. For example, the original weight of the color feature is 0.3. Since its prediction uncertainty is relatively high, the weight needs to be reduced by 0.1905×0.3 ≈ 0.057. Then the adjusted weight of the color feature is 0.3 - 0.057 ≈ 0.243, that is, the third weight information is 0.243. At the same time, adjust the weights of other texture features according to the corresponding strategy, and ensure that the sum of the adjusted weights is still 1.

[0091] Based on the third texture data and the fourth texture data, the respective basic similarities corresponding to each texture feature can be calculated, that is, the basic similarity S1 of the color feature, the basic similarity S2 of the shape feature, and the basic similarity S3 of the distribution feature.

[0092] Based on the preset seventh formula and each basic similarity, the second sub-texture similarity can be calculated. The seventh formula can be expressed as:

[0093] where, represents the second sub-texture similarity; represents the third weight information corresponding to the color feature; represents the third weight information corresponding to the shape feature; represents the third weight information corresponding to the distribution feature.

[0094] By considering the degree of change of each texture feature during the process of playing with the object, the weights are dynamically adjusted, so that the calculation of the second sub-texture similarity can better reflect the actual texture change. Texture features with large changes may have a greater impact on the second sub-texture similarity, so higher weights can be given.

[0095] The prediction uncertainty reflects the credibility of the model's prediction result for a certain texture feature. If the prediction uncertainty of a certain texture feature is high, it means that the model's prediction of it is not accurate enough. By reducing the weight of this texture feature, the impact of the prediction error on the calculation of the second sub-texture similarity can be reduced, and the reliability of the calculation of the second sub-texture similarity can be improved.

[0096] Based on the same technical concept, the present application also provides a pairing analysis device for paired nuclear objects, as Figure 5 shown. The pairing analysis device 200 for paired nuclear objects mainly includes: The contour extraction module 201 is used to, for each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations; The single contour similarity acquisition module 202 is used to, for each target paired object at the current orientation, based on the first contour data corresponding to the current orientation and the second contour data corresponding to the current orientation, perform center point alignment processing, and calculate the single contour similarity between the current paired object and the target paired object in the current orientation; The average contour similarity acquisition module 203 is used to, for each target paired object, based on the single contour similarities corresponding to different orientations respectively, obtain the average contour similarity between the current paired object and the target paired object; The texture similarity acquisition module 204 is used to, based on the first image data, extract the first texture data corresponding to the current paired object, for each target paired object, based on the second image data, extract the second texture data corresponding to the target paired object, and based on the first texture data and the second texture data, obtain the texture similarity between the current paired object and the target paired object; The comprehensive analysis module 205 is used to, for each target paired object, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object.

[0097] Optionally, before the contour extraction module 201, it further includes: The first acquisition and screening module is used to obtain first screening information corresponding to the current paired object based on the first image data, where the first screening information includes first size information and first color information; The second acquisition and screening module is used to obtain second screening information corresponding to each of the multiple candidate paired objects based on the second image data corresponding to each of the multiple candidate paired objects, and each second screening information includes second size information and second color information; The screening and processing module is used to screen the multiple candidate paired objects based on the first screening information and the multiple second screening information, and use the screened candidate paired objects as the target paired objects respectively.

[0098] Optionally, before the comprehensive analysis module 205, it further includes: The first smoothness calculation module is used to obtain a first area value and a first gray value corresponding to the first illuminated area based on the third image data corresponding to the current paired object, and calculate the first surface smoothness corresponding to the current paired object based on the first area value and the first gray value; the third image data is the image data collected under the illumination of an auxiliary light source at a preset angle; The second smoothness calculation module is used to, for each of the target paired objects, obtain a second area value and a second gray value corresponding to the second illuminated area based on the fourth image data corresponding to the target paired object, and calculate the second surface smoothness corresponding to the target paired object based on the second area value and the second gray value; the fourth image data is the image data collected under the illumination of the auxiliary light source; The gloss similarity calculation module is used to, for each of the target paired objects, calculate the gloss similarity between the current paired object and the target paired object based on the first surface smoothness and the second surface smoothness.

[0099] Optionally, the comprehensive analysis module 205 includes: The weight acquisition sub-module is used to obtain first weight information corresponding to the average contour similarity, the texture similarity, and the gloss similarity respectively based on the target requirements and attribute information corresponding to the current paired object, where the target requirements are parameters representing the user's consistency requirements for paired objects, and the attribute information is parameters representing the physical characteristics of paired objects; The weighted calculation sub-module is used to perform weighted calculation on the average contour similarity, the texture similarity, and the gloss similarity based on the respective first weight information to obtain the comprehensive similarity information, and obtain the pairing analysis result based on the comprehensive similarity information, where the pairing analysis result includes pairing success and pairing failure.

[0100] Optionally, before the contour extraction module 201, it further includes: The first angle acquisition module is configured to acquire first current angle information of the current paired object in a preset coordinate system based on the fifth image data corresponding to the current paired object; The second angle acquisition module is configured to, for each of the target paired objects, acquire second current angle information of the target paired object in a preset coordinate system based on the sixth image data corresponding to the target paired object; The processing execution module is configured to execute the process of acquiring the first image data when the first current angle information meets the preset first target angle information; and execute the process of acquiring the second image data when the second current angle information meets the preset second target angle information.

[0101] Optionally, for each of the target paired objects, the texture similarity acquisition module 204 includes: The first calculation sub-module is configured to, for each of the target paired objects, perform an overlapping process on the first texture data and the second texture data, acquire the intersection texture area and the union texture area between the first texture data and the second texture data, and calculate a first sub-texture similarity between the current paired object and the target paired object based on the intersection texture area and the union texture area; The prediction sub-module is configured to, for each of the target paired objects, predict third texture data corresponding to the current paired object and fourth texture data corresponding to the target paired object based on a trained texture change prediction model, the first texture data, and the second texture data, where the texture data includes color features, shape features, and distribution features; The second calculation sub-module is configured to, for each of the target paired objects, calculate a second sub-texture similarity between the current paired object and the target paired object based on the third texture data and the fourth texture data; The comprehensive acquisition sub-module is configured to, for each of the target paired objects, acquire the texture similarity based on the first sub-texture similarity and the second sub-texture similarity.

[0102] Optionally, the second calculation sub-module includes: The first determination sub-module is configured to determine the respective change degrees of the color features, shape features, and distribution features in the texture data based on the first texture data, the second texture data, the third texture data, and the fourth texture data; A second determination sub-module, configured to respectively determine second weight information corresponding to the color feature, the shape feature, and the distribution feature based on each of the degrees of change and a preset weight adjustment rule; the weight adjustment rule includes that within the weight range corresponding to each feature, the greater the degree of change of the feature, the higher the corresponding second weight information; An uncertainty acquisition sub-module, configured to acquire prediction uncertainties for the color feature, the shape feature, and the distribution feature based on the texture change prediction model; A comprehensive calculation sub-module, configured to respectively adjust the second weight information corresponding to the color feature, the shape feature, and the distribution feature based on each of the prediction uncertainties to obtain third weight information, and calculate the second sub-texture similarity based on the third weight information.

[0103] Those skilled in the art can clearly understand that for the sake of convenience and brevity of description, the specific working processes of the above-described system, apparatus, and module can refer to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0104] Based on the same technical concept, the present application further provides an electronic device, as Figure 6 shown. The electronic device 300 includes a processor 301 and a memory 302, and may further include one or more of an information input / output (I / O) interface 303, a communication component 304, and a communication bus 305.

[0105] Among them, the processor 301 is used to control the overall operation of the electronic device 300 to complete all or part of the steps in the above-mentioned pairing analysis method for paired combined nuclear objects; the memory 302 is used to store various types of data to support the operation of the electronic device 300. These data may include, for example, instructions for any application program or method operating on the electronic device 300, and application program-related data. The memory 302 may be implemented by any type of volatile or non-volatile storage device or a combination thereof.

[0106] The I / O interface 303 provides an interface between the processor 301 and other interface modules. The above-mentioned other interface modules may be a keyboard, a mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 304 is used to test wired or wireless communication between the electronic device 300 and other devices.

[0107] The communication bus 305 may include a path for transmitting information among the above components. The communication bus 305 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The communication bus 305 may be divided into an address bus, a data bus, a control bus, etc.

[0108] The electronic device 300 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components, and is used to execute the pairing analysis method of the nuclear objects in paired combinations given in the above embodiments.

[0109] The electronic device 300 may include, but is not limited to, mobile terminals such as digital broadcast receivers, PDAs (Personal Digital Assistants), PMPs (Portable Multimedia Players), etc., and fixed terminals such as digital TVs, desktop computers, etc., and may also be a server, etc.

[0110] Based on the same technical concept, the present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the pairing analysis method of the nuclear objects in paired combinations described above are implemented.

[0111] The computer-readable storage medium may include various media capable of storing program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0112] The term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0113] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present application, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0114] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0115] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A pairing analysis method for a pair of combined nuclear objects, characterized in that, Including: For each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations; For each target paired object at the current orientation, based on the first contour data corresponding to the current orientation and the second contour data corresponding to the current orientation, perform center point alignment processing, and calculate the single contour similarity between the current paired object and the target paired object in the current orientation; For each target paired object, based on the single contour similarities corresponding to different orientations respectively, obtain the average contour similarity between the current paired object and the target paired object; Based on the first image data, extract the first texture data corresponding to the current paired object. For each target paired object, based on the second image data, extract the second texture data corresponding to the target paired object. Based on the first texture data and the second texture data, obtain the texture similarity between the current paired object and the target paired object; For each target paired object, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object.

2. The pairing analysis method of a pair of combined nuclear objects according to claim 1, characterized in that, Before the step of "For each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object at different orientations and the second contour data corresponding to the target paired object at different orientations", it further includes: Based on the first image data, obtain the first screening information corresponding to the current paired object, and the first screening information includes the first size information and the first color information; Based on the second image data corresponding to each of the multiple candidate paired objects, obtain the second screening information corresponding to each of the candidate paired objects, and each second screening information includes the second size information and the second color information; Based on the first screening information and the multiple second screening information, screen the multiple candidate paired objects, and use the screened candidate paired objects as the target paired objects respectively.

3. The paired analysis method of a paired combination of nuclear objects according to claim 1 or 2, characterized in that, Before the step of "For each target paired object, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object", it further includes: Based on the third image data corresponding to the current paired object, obtain the first area value and the first gray value corresponding to the first illuminated area, and based on the first area value and the first gray value, calculate the first surface smoothness degree of the current paired object; the third image data is the image data collected under the irradiation of an auxiliary light source at a preset angle. For each of the target paired objects, based on the fourth image data corresponding to the target paired object, obtain the second area value and the second gray value corresponding to the second illuminated area, and calculate the second surface smoothness corresponding to the target paired object based on the second area value and the second gray value; the fourth image data is the image data collected based on the illumination of the auxiliary light source. For each of the target paired objects, calculate the gloss similarity between the current paired object and the target paired object based on the first surface smoothness and the second surface smoothness.

4. The pairing analysis method of a pair of combined nuclear objects according to claim 3, characterized in that, For each of the target paired objects, based on the average contour similarity and the texture similarity, obtain the comprehensive similarity information and the pairing analysis result between the current paired object and the target paired object, including: Based on the target requirements and attribute information corresponding to the current paired object, obtain the first weight information corresponding to the average contour similarity, the texture similarity, and the gloss similarity respectively. The target requirements are parameters representing the user's consistency requirements for paired objects, and the attribute information is parameters representing the physical characteristics of paired objects. Based on each of the first weight information, perform a weighted calculation on the average contour similarity, the texture similarity, and the gloss similarity to obtain the comprehensive similarity information, and obtain the pairing analysis result based on the comprehensive similarity information. The pairing analysis result includes pairing success and pairing failure.

5. The pairing analysis method of a pair of combined nuclear objects according to claim 1, characterized in that, Before, for each target paired object, based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object, extract the first contour data corresponding to the current paired object in different orientations and the second contour data corresponding to the target paired object in different orientations, further including: Based on the fifth image data corresponding to the current paired object, obtain the first current angle information of the current paired object in the preset coordinate system. For each of the target paired objects, based on the sixth image data corresponding to the target paired object, obtain the second current angle information of the target paired object in the preset coordinate system. If the first current angle information meets the preset first target angle information, then execute the step of collecting the first image data; if the second current angle information meets the preset second target angle information, then execute the step of collecting the second image data.

6. The pairing analysis method of a pair of combined nuclear objects according to claim 1, characterized in that, For each of the target paired objects, the obtaining of the texture similarity between the current paired object and the target paired object based on the first texture data and the second texture data includes: For each of the target paired objects, perform an overlapping process on the first texture data and the second texture data, obtain the intersection texture area and the union texture area between the first texture data and the second texture data, and calculate the first sub-texture similarity between the current paired object and the target paired object based on the intersection texture area and the union texture area. For each of the target paired objects, based on the trained texture change prediction model, the first texture data, and the second texture data, predict the third texture data corresponding to the current paired object and the fourth texture data corresponding to the target paired object, where the texture data includes color features, shape features, and distribution features; For each of the target paired objects, based on the third texture data and the fourth texture data, calculate the second sub-texture similarity between the current paired object and the target paired object; For each of the target paired objects, based on the first sub-texture similarity and the second sub-texture similarity, obtain the texture similarity.

7. The pairing analysis method of a pair of combined nuclear objects according to claim 6, characterized in that, The step of, for each of the target paired objects, calculating the second sub-texture similarity between the current paired object and the target paired object based on the third texture data and the fourth texture data includes: Based on the first texture data, the second texture data, the third texture data, and the fourth texture data, determine the respective degrees of change of the color features, shape features, and distribution features in the texture data; Based on each of the degrees of change and a preset weight adjustment rule, respectively determine the second weight information corresponding to the color features, the shape features, and the distribution features; the weight adjustment rule includes that within the weight range corresponding to each feature, the greater the degree of change of the feature, the higher the corresponding second weight information; Based on the texture change prediction model, obtain the prediction uncertainties of the color features, the shape features, and the distribution features; Based on each of the prediction uncertainties, respectively adjust the second weight information corresponding to the color features, the shape features, and the distribution features to obtain the third weight information, and based on the third weight information, calculate the second sub-texture similarity.

8. A paired analysis device for paired nuclear objects, characterized in that, Includes: An extraction contour module, which is used for each target paired object to extract the first contour data corresponding to the current paired object in different orientations and the second contour data corresponding to the target paired object in different orientations based on the first image data corresponding to the current paired object and the second image data corresponding to the target paired object; A single contour similarity acquisition module, which is used for each target paired object in the current orientation to perform center point alignment processing based on the first contour data corresponding to the current orientation and the second contour data corresponding to the current orientation, and calculate the single contour similarity between the current paired object and the target paired object in the current orientation; An average contour similarity acquisition module, which is used for each target paired object to obtain the average contour similarity between the current paired object and the target paired object based on the single contour similarities corresponding to different orientations; A texture similarity acquisition module, configured to extract first texture data corresponding to the current paired object based on the first image data, and for each of the target paired objects, extract second texture data corresponding to the target paired object based on the second image data, and acquire the texture similarity between the current paired object and the target paired object based on the first texture data and the second texture data; A comprehensive analysis module, configured to, for each of the target paired objects, acquire comprehensive similarity information and paired analysis results between the current paired object and the target paired object based on the average contour similarity and the texture similarity.

9. An electronic device, characterized in that, Comprising a processor and a memory, the processor being coupled to the memory; The processor is configured to execute a computer program stored in the memory, so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Comprising a computer program or instruction, when the computer program or instruction runs on a computer, the computer is caused to execute the method according to any one of claims 1 to 7.

Images