Dynamic matching recognition method and system for multi-modal trace morphological features

Abstract

The application discloses a multi-modal trace morphology feature dynamic matching recognition method and system, relates to the technical field of feature matching, and comprises the following steps: color feature extraction is performed on a trace image to obtain a color feature value; a matching reference color feature value set is compared with the color feature value to obtain a color similarity set; it is judged whether the color coefficient meets a predetermined coefficient threshold value; if yes, a texture feature value set is obtained; a first texture feature value set corresponding to a first tool in a tool feature database is extracted, and the texture feature value set is compared to obtain a first texture similarity; when the similarity reaches a predetermined similarity limit value, a first type corresponding to the first tool is taken as a tool type of the trace image. The application solves the technical problems that, in the prior art, trace images are interfered by factors such as material, illumination, corrosion or thermal discoloration, leading to poor recognition accuracy and the inability to stably determine the tool type, and achieves the technical effects of improving tool trace recognition accuracy and stability under complex conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of feature matching, and in particular to a dynamic matching and recognition method and system for multimodal trace morphological features. Background Art

[0002] The types of traces left by tools on trace-bearing objects made of different materials, such as metal, plastic, and fabric, vary greatly, including shoe tread imprints, tool slip marks, and tire tread patterns. These traces exhibit significant structural differences in images and are susceptible to factors such as material reflectivity, lighting changes, corrosion and oxidation, and high-temperature discoloration, resulting in image color distortion, detail loss, or blurred textures, which seriously interfere with the stable extraction of trace features. In particular, in actual acquisition scenarios, trace areas often suffer from local occlusion, surface contamination, or color drift, further reducing the separability and contrast of color and texture features. Traditional recognition methods that rely on a single visual feature have difficulty maintaining recognition accuracy and consistency in such complex environments, limiting the ability to stably determine and automatically identify tool types. Summary of the Invention

[0003] The present application provides a dynamic matching and recognition method and system for multimodal trace morphological features, which is used to solve the technical problems in the prior art such as poor recognition accuracy and inability to stably determine the tool type due to the interference of trace images with factors such as material, lighting, corrosion or thermal discoloration.

[0004] In view of the above problems, the present application provides a dynamic matching and recognition method and system for multimodal trace morphological features.

[0005] In a first aspect of the present application, a method for dynamic matching and recognition of multimodal trace morphological features is provided, the method comprising:

[0006] Perform color feature extraction on the collected trace image to obtain color feature values, wherein the trace image refers to an image of the trace of use of the tool on the trace-bearing object; match the baseline color feature value set of the trace-bearing object, and compare the color feature value with the baseline color feature value set to obtain a color similarity set; judge and analyze whether the color coefficient obtained by analyzing the color similarity set meets the predetermined coefficient threshold; if it meets, issue a texture feature collection instruction, and obtain the texture feature value set of the trace image according to the texture feature collection instruction; extract the first texture feature value set corresponding to the first tool in the tool feature database; compare the texture feature value set with the first texture feature value set to obtain a first texture similarity; when the first texture similarity reaches a predetermined similarity limit, use the first type corresponding to the first tool as the tool type of the trace image.

[0007] A second aspect of the present application provides a dynamic matching and recognition system for multimodal trace morphological features, the system comprising:

[0008] a feature extraction module for extracting color features from a collected trace image to obtain a color feature value, wherein the trace image refers to an image of the trace of use of a tool on a trace-bearing object; a matching and comparison module for matching a reference color feature value set of the trace-bearing object, and comparing the color feature value with the reference color feature value set to obtain a color similarity set; a judgment module for judging whether a color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; a texture feature collection module for issuing a texture feature collection instruction if it meets the requirements, and obtaining a texture feature value set of the trace image according to the texture feature collection instruction; an extraction module for extracting a first texture feature value set corresponding to a first tool in a tool feature database; a similarity determination module for comparing the texture feature value set with the first texture feature value set to obtain a first texture similarity; a tool type confirmation module for using a first type corresponding to the first tool as the tool type of the trace image when the first texture similarity reaches a predetermined similarity limit.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] The present application extracts color features from a collected trace image to obtain a color feature value, wherein the trace image refers to an image of a tool's use trace on a trace-bearing object; matches the reference color feature value set of the trace-bearing object, and compares the color feature value with the reference color feature value set to obtain a color similarity set; determines whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; if so, issues a texture feature collection instruction, and obtains a texture feature value set of the trace image according to the texture feature collection instruction; extracts a first texture feature value set corresponding to a first tool in a tool feature database; compares the texture feature value set with the first texture feature value set to obtain a first texture similarity; when the first texture similarity reaches a predetermined similarity limit, the first type corresponding to the first tool is used as the tool type of the trace image. The present invention solves the technical problem in the prior art that the trace image is interfered with by factors such as material, lighting, corrosion, or thermal discoloration, resulting in poor recognition accuracy and inability to stably determine the tool type. By introducing a dual feature fusion and discrimination mechanism of color and texture, the trace image is dynamically matched with the tool template feature, achieving the technical effect of improving the accuracy and stability of tool trace recognition under complex conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic flow chart of a method for dynamic matching and identification of multimodal trace morphological features provided in an embodiment of the present application;

[0013] Figure 2 Schematic diagram of the structure of the dynamic matching and recognition system for multimodal trace morphological features provided in an embodiment of the present application.

[0014] Explanation of reference numerals: feature extraction module 11 , matching and comparison module 12 , judgment module 13 , texture feature collection module 14 , extraction module 15 , similarity determination module 16 , tool type confirmation module 17 . DETAILED DESCRIPTION

[0015] This application provides a dynamic matching and recognition method and system for multimodal trace morphological features, aiming to solve the technical problems in the prior art caused by the interference of trace images with factors such as material, lighting, corrosion or thermal discoloration, resulting in poor recognition accuracy and inability to stably determine the type of tool. By introducing a dual feature fusion and discrimination mechanism of color and texture, the trace image and tool template features are dynamically matched, achieving the technical effect of improving the accuracy and stability of tool trace recognition under complex conditions.

[0016] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0017] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0018] Example 1, as Figure 1 As shown, the present application provides a dynamic matching and recognition method for multimodal trace morphological features, the method comprising:

[0019] Step S100: performing color feature extraction on the collected trace image to obtain a color feature value, wherein the trace image refers to an image of the trace of use of the tool on the trace-bearing object.

[0020] In the embodiments of this application, a trace image refers to image information containing indentations, patterns, scratches, or structural textures formed by a tool on a trace-bearing object (such as a metal surface, rubber product, or textile material). This includes forms such as shoe tread marks, striped friction marks, or surface impressions produced by the edges of tool structures. When capturing such trace images, a high-resolution industrial camera or microscopic image acquisition device is used to capture the target trace area under uniform lighting conditions and background environment to avoid image color casts caused by surface reflections, material absorption differences, and other factors, ensuring that the captured image has good color consistency and detail retention.

[0021] Then the color feature extraction is performed on the collected trace image by calculating the discrete cosine transform (DCT) coefficients of the image and extracting the DC coefficient as the color feature value.

[0022] Furthermore, in the method provided in the embodiment of the application, color feature extraction is performed on the collected trace image to obtain color feature values, and the method further includes:

[0023] Obtaining discrete cosine transform coefficients of the trace image; and using a DC coefficient in the discrete cosine transform coefficients to represent the color feature value.

[0024] In this embodiment, the captured trace image is first preprocessed, including grayscale conversion to reduce channel complexity while preserving overall brightness information. A two-dimensional discrete cosine transform (2D-DCT) is then applied to the entire grayscale image, converting the image from the spatial domain to the frequency domain. This yields a set of discrete cosine transform coefficients. These coefficients form a coefficient matrix with the same size as the original image, where each coefficient corresponds to the energy distribution of a different spatial frequency component.

[0025] After the transformation is completed, the DC coefficient (DC coefficient) at the upper left corner is extracted from the resulting DCT coefficient matrix. This coefficient is a concentrated expression of the overall color or brightness level of the image. Ultimately, this DC coefficient is used as the color feature value of the trace image.

[0026] Step S200: matching a reference color feature value set of the mark-bearing object, and comparing the color feature value with the reference color feature value set to obtain a color similarity set.

[0027] In this embodiment, the process of matching a scar-bearing object's reference color feature value set first constructs color reference data for the scar-bearing object in various states. Specifically, this involves obtaining the scar-bearing object's original image, a set of corroded images, and a set of high-temperature images, and extracting the original color feature values, corroded color feature value sets, and high-temperature color feature value sets from each image. Together, these three sets constitute the reference color feature value set.

[0028] Subsequently, the color feature values ​​extracted from the collected trace image are compared with the various color feature values ​​in the benchmark set in turn, and the similarity index is calculated using methods such as Euclidean distance, thereby obtaining a set of results reflecting the degree of color similarity, namely the color similarity set.

[0029] Furthermore, in the method provided in the embodiment of the application, matching the reference color feature value set of the mark-bearing object further includes:

[0030] Acquire the original image of the scar-bearing object and obtain the original color feature value of the original image; establish a corrosion image set of the scar-bearing object and sequentially obtain the corrosion color feature value set of each corrosion image in the corrosion image set; establish a high-temperature image set of the scar-bearing object and sequentially obtain the high-temperature color feature value set of each high-temperature image in the high-temperature image set; establish the reference color feature value set based on the original color feature value, the corrosion color feature value set and the high-temperature color feature value set.

[0031] In the embodiments of the present application, to obtain the original image of the scarred object, image acquisition is first performed under standardized shooting conditions. Specifically, an image acquisition device with high-resolution imaging capabilities (such as an industrial camera or a microscope) is used to image the surface of the scarred object, unaffected by external physical and chemical influences, in a controlled environment with constant lighting, no stray reflections, and a uniform background. The original image is captured in its true color state. After image acquisition, the color feature extraction phase begins. During color feature extraction, the entire image is used as input. First, the image is grayscaled to preserve its overall brightness information and reduce channel complexity. A two-dimensional discrete cosine transform (DCT) is then performed on the entire grayscale image to convert the image from the spatial domain to the frequency domain, thereby obtaining the image's energy distribution across different frequency components. In the resulting DCT coefficient matrix, the DC coefficient in the upper left corner reflects the average brightness level or color trend of the entire image. This DC coefficient is directly extracted and used as the original color feature value of the original image.

[0032] Next, a controlled corrosion operation is applied to the surface of the object bearing the mark to simulate common processes in the natural environment, such as oxidation, acid etching, or wear. Corrosion methods can include chemical corrosion (such as dilute acid treatment) or physical corrosion (such as surface micro-abrasion). Images of the surface of the object bearing the mark are collected at different corrosion stages or time points, forming a set of corrosion images reflecting different corrosion states. Each image in the corrosion image set is grayscaled and subjected to the same DCT processing as the original image. The DC coefficient of the image is extracted as the corrosion color feature value of the image. Finally, the color feature values ​​under multiple corrosion states are combined to form a corrosion color feature value set.

[0033] The scar-bearing object then undergoes a controlled high-temperature treatment, heating it to multiple set temperatures (e.g., 100°C, 150°C, and 200°C) using infrared heaters, electric hot plates, and other methods. Surface images are captured while the thermal field is stable, generating a set of high-temperature images. The aforementioned process is repeated for each high-temperature image, extracting its overall DCT DC coefficient as a high-temperature color feature value. The color feature results from all high-temperature images are then aggregated to form a set of high-temperature color feature values.

[0034] Finally, the original color eigenvalues ​​of the original image, the eroded color eigenvalue set of the eroded image set, and the high temperature color eigenvalue set of the high temperature image set are unified and sorted to obtain a reference color eigenvalue set.

[0035] Furthermore, in the method provided in the embodiment of the application, before determining whether the color coefficient obtained by analyzing the color similarity set meets the predetermined coefficient threshold, the method further includes:

[0036] Extracting a first color similarity from the color similarity set, wherein the first color similarity refers to the similarity between the color feature value and the original color feature value; extracting a second color similarity from the color similarity set, wherein the second color similarity refers to the color similarity between the trace image and a target corrosion image in the corrosion image set, and the target corrosion image refers to the corrosion image with the highest similarity to the color feature value; extracting a third color similarity from the color similarity set, wherein the third color similarity refers to the color similarity between the trace image and a target high-temperature image in the high-temperature image set, and the target high-temperature image refers to the high-temperature image with the highest similarity to the color feature value; obtaining the color coefficient according to the first color similarity, the second color similarity, and the third color similarity.

[0037] In this embodiment of the present application, a first color similarity is first extracted. The first color similarity refers to the similarity between the color feature value and the original color feature value. By calculating the Euclidean distance between these two feature values, a numerical value representing their color deviation is obtained, which serves as the original evaluation value of the first color similarity.

[0038] Next, the second color similarity is extracted. The second color similarity refers to the color similarity between the trace image and the target corrosion image in the corrosion image set. First, the Euclidean distance is calculated between the color feature values ​​of the trace image and the corresponding corrosion color feature values ​​of each image in the corrosion image set. The image with the minimum distance is found. This image is the target corrosion image. The corresponding minimum distance value reflects the degree of fit between the color features of the corrosion state and the trace image, which is the second color similarity.

[0039] Next, the third color similarity is extracted. This refers to the color similarity between the trace image and the target high-temperature image in the high-temperature image set. The process is the same as for the second color similarity. The Euclidean distance between the color feature values ​​of the trace image and the high-temperature color feature values ​​of each high-temperature image is calculated. The image with the minimum distance is the target high-temperature image. This Euclidean distance is the third color similarity, reflecting whether the trace image exhibits color characteristics under high temperature.

[0040] Finally, the color coefficient is calculated based on the extracted first, second, and third color similarities. Specifically, the second and third color similarities are first averaged to obtain a first average. This first average is then arithmetic averaged with the first color similarity to form a fusion evaluation value. Finally, this fusion value is normalized to output the final color coefficient.

[0041] Furthermore, in the method provided in the embodiment of the application, obtaining the color coefficient according to the first color similarity, the second color similarity, and the third color similarity further includes:

[0042] The average of the second color similarity and the third color similarity is obtained, recorded as a first average; the average of the first average and the first color similarity is obtained, and normalized to obtain the color coefficient.

[0043] In the embodiment of the present application, an arithmetic mean operation is first performed on the second color similarity and the third color similarity to obtain a first mean value.

[0044] Next, the first mean and the first color similarity are averaged to form a fusion evaluation value. Finally, the fusion evaluation value is normalized to a uniform value range based on the set theoretical maximum and minimum value ranges to generate the final color coefficient.

[0045] Furthermore, in the method provided in the embodiment of the application, before determining whether the color coefficient obtained by analyzing the color similarity set meets the predetermined coefficient threshold, the method further includes:

[0046] Reading predetermined object features, and collecting multi-dimensional features of the scar-bearing object based on the predetermined object features to obtain an object feature set; performing normalized weighted analysis on the object feature set to obtain an object performance feedback coefficient; and adjusting the color coefficient using the object performance feedback coefficient as a weight.

[0047] In this embodiment, the first step is to read predetermined object features. This feature information, pre-stored in the device database, primarily includes structural and physical properties that significantly influence trace formation, such as material type (e.g., metal, plastic), hardness grade (e.g., Brinell HB value), surface condition (e.g., roughness Ra value, coating), geometry (e.g., flat, cylindrical), and thickness. These features are loaded via identification numbers, user input, or tag recognition, serving as reference standards to form a predetermined feature template.

[0048] Then, based on the predetermined object characteristics, multi-dimensional features of the object bearing the marks are collected. This process uses direct measurement methods, using integrated sensors or detection devices to measure the object in real-world conditions. For example, a surface roughness meter is used to obtain Ra values, an ultrasonic thickness gauge is used to obtain thickness data, and image recognition is used to analyze the object's geometric structure. The various measurement results are sequentially populated according to the predetermined feature items to form an object feature set consistent with the template structure.

[0049] Next, a normalized weighted analysis is performed on the aforementioned object feature set to obtain the object performance feedback coefficient. Normalization utilizes a minimum-maximum normalization method, linearly scaling the measured values ​​to the [0, 1] interval based on the physically feasible range of each feature (e.g., metal hardness is typically between 50 and 300 HB). Subsequently, a fixed-weight weighting method is used to assign a preset weight to each normalized feature, such as 0.35 for hardness, 0.25 for surface condition, 0.2 for shape and structure, and 0.2 for thickness. By multiplying the normalized feature values ​​by the corresponding weights and summing the results, a value between 0 and 1 is obtained, which is the object performance feedback coefficient.

[0050] Finally, the color coefficient is adjusted using the object performance feedback coefficient as a weight. This step uses a linear multiplication method, that is, the color coefficient obtained in the previous step is multiplied by the object performance feedback coefficient to generate the adjusted color coefficient.

[0051] Step S300: determining whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold.

[0052] In the embodiment of the present application, after adjusting the color coefficients obtained from the color similarity set, the adjusted color coefficients are compared with a predetermined coefficient threshold set by a technical expert to determine whether they meet the predetermined coefficient threshold. If the color coefficient is greater than or equal to the predetermined coefficient threshold, it is considered to meet the requirements. If it is less than the threshold, it is considered to be non-compliant.

[0053] Furthermore, in the method provided in the embodiment of the application, after determining whether the color coefficient obtained by analyzing the color similarity set meets the predetermined coefficient threshold, the method further includes:

[0054] If it does not meet the requirements, a repair instruction is issued; based on the repair instruction, an image repair plan is retrieved to repair the trace image; wherein, the image repair plan includes a physical repair plan and / or a chemical repair plan, and the physical repair plan includes plasma cleaning and laser cleaning, and the chemical repair plan includes composite acid cleaning and oxidation-reduction.

[0055] In this embodiment of the present application, when it is determined that the color coefficient does not reach the predetermined coefficient threshold, a repair instruction is issued. This repair instruction serves as a process control signal to call and execute a preset image repair plan to improve the clarity and color accuracy of the trace area in the image, thereby meeting subsequent recognition requirements.

[0056] Image restoration plans are categorized into physical and / or chemical restoration plans based on the type of image contamination or interference. These plans automatically match execution paths based on the material, contamination form, and cause of the image. In physical restoration plans, if image quality degradation is due to physical interference such as surface dust, particle adhesion, or weakly adhered coatings, plasma cleaning is prioritized. This method utilizes high-energy active particles in plasma (such as oxygen or argon ions) to perform low-temperature, non-destructive surface treatment on the traced area, removing micron-sized contaminants while preserving the trace's texture characteristics.

[0057] If the contamination is more stable coating residue, oxidation plaque, or aged film, laser cleaning is used. This method precisely removes surface contaminants by adjusting the power, wavelength, and focus of the pulsed laser beam, utilizing photothermal or photolytic effects.

[0058] In the chemical restoration plan, if the trace image surface contains an oxide film, chemical reaction deposits, or color migration interference, a composite pickling solution is used. Composite pickling uses multiple acids (such as a nitric acid-phosphoric acid mixture) to react with the contaminants to remove the metal oxide layer and deposits.

[0059] If the image is affected by heat or contaminated by chemical reducing agents, redox treatment is used. This method involves applying an oxidizing agent (such as hydrogen peroxide) or a reducing agent (such as sulfite) to the contaminated layer, causing a redox reaction, which then changes its structure and is then removed. This method is commonly used for removing spots or restoring color layers.

[0060] After completing the physical or chemical restoration, recapture the image and re-execute the color feature extraction and color coefficient calculation process to ensure that the image quality meets the recognition requirements.

[0061] Step S400: If it is consistent, a texture feature collection instruction is issued, and a texture feature value set of the trace image is obtained according to the texture feature collection instruction.

[0062] In this embodiment of the present application, when a color coefficient is determined to meet a predetermined coefficient threshold, a texture feature collection instruction is issued, and the texture information extraction process for the trace image begins. Specifically, the trace image is first rasterized, dividing the image into multiple grid blocks. Subsequently, these grid blocks are uniformly sampled based on a predetermined sampling strategy to obtain a representative set of target sample blocks. During the extraction process, the first sample block is selected and a discrete cosine transform (DCT) is performed on it. The first alternating current coefficient (AC coefficient) is extracted from it as the first texture feature value of the sample block. This feature value is then used to construct a complete set of texture feature values.

[0063] Furthermore, in the method provided in the embodiment of the application, obtaining the texture feature value set of the trace image according to the texture feature collection instruction further includes:

[0064] The trace image is rasterized to obtain a grid block set; the grid block set is uniformly sampled based on a predetermined sampling strategy to obtain a target sample block set; a first sample block in the target sample block set is extracted, and a first AC coefficient in a first discrete cosine transform coefficient of the first sample block is used as a first texture feature value; and the texture feature value set is assembled based on the first texture feature value according to the texture feature collection instruction.

[0065] In the embodiment of the present application, rasterization processing is first performed on the trace image, that is, the entire image is divided into multiple adjacent non-overlapping image blocks of a fixed size (such as 8×8 pixels), which are called raster block sets.

[0066] Next, a subset of the grid block set is selected based on the built-in uniform sampling strategy. This strategy uses a uniform sampling method with equal spacing between rows and columns. That is, spatially representative image blocks are extracted from the entire grid block set at a set step size (for example, selecting every other block) to form the target sample block set.

[0067] The first image block from the target sample block set is then selected as the first sample block, and a two-dimensional discrete cosine transform (2D-DCT) is performed on this block. The DCT calculation is performed on this image block, generating a corresponding frequency coefficient matrix. The coefficient in the upper left corner of this matrix is ​​the DC coefficient (DC), which represents the average brightness of the image block. The first AC coefficient (AC) immediately adjacent to the DC coefficient, i.e., the first frequency component in the lowest non-zero frequency direction in the matrix, reflects the underlying texture variation of the image block in the horizontal or vertical direction.

[0068] The first AC coefficient is extracted as the first texture feature value, which is used to characterize the basic texture structure of the image block. This operation is repeated for the entire target sample block set, extracting the first AC coefficient for each sample block and aggregating the texture feature values ​​of all sample blocks to form a complete texture feature value set.

[0069] Step S500: extracting a first texture feature value set corresponding to a first tool from a tool feature database.

[0070] In this embodiment of the present application, a first texture feature value set corresponding to a first tool is extracted from a tool feature database. The first tool refers to a tool object selected in a preset order during the current recognition process as one of the candidate matching tools. The first texture feature value set refers to a set of texture feature data extracted through a specific processing flow from a trace image formed by the tool acting on a trace-bearing object under standard conditions.

[0071] The tool feature database pre-stores image samples of typical traces produced by various types of tools on standard contact surfaces. These traces include shoe tread imprints, strip marks left by tool edge sliding, and local image structures formed by specific pattern outlines. During the database construction phase, these images undergo a unified processing pipeline, including image rasterization, extraction of sample image blocks based on a uniform sampling strategy, and performing a two-dimensional discrete cosine transform (2D-DCT) on each image block. The first representative alternating current coefficient (AC coefficient) is then extracted to capture local frequency characteristics. During this step, the database entry for the standard trace image corresponding to the first tool is retrieved and its stored AC coefficient set is directly read. These AC coefficients reflect the spatial frequency variations of the trace image across multiple regions and constitute the first set of texture feature values.

[0072] Step S600: Compare the texture feature value set with the first texture feature value set to obtain a first texture similarity.

[0073] In an embodiment of the present application, the texture feature value set of the trace image and the first texture feature value set of the first tool are sequentially curve-processed to construct corresponding texture curves and first texture curves, respectively; then, the first Fletcher distance between the two texture curves is used as a measurement indicator to calculate the similarity between the two in morphological trends, thereby obtaining the final first texture similarity.

[0074] Furthermore, in the method provided in the embodiment of the application, comparing the texture feature value set with the first texture feature value set to obtain the first texture similarity further includes:

[0075] The texture feature value set and the first texture feature value set are sequentially subjected to curve processing to obtain a texture curve and a first texture curve respectively; and the first texture similarity is represented by a first Fletcher distance between the texture curve and the first texture curve.

[0076] In an embodiment of the present application, the texture feature value set obtained by extracting the trace image is first processed, the feature values ​​in the set are arranged in sequence according to the order of image block extraction to form a one-dimensional series, and the sequence is continuous processed using a linear interpolation method to generate a texture curve.

[0077] The same linear interpolation operation is performed on the first texture feature value set corresponding to the first tool. The feature values ​​in the set are arranged in a fixed order and a continuous curve is constructed using the same interpolation method to form a standard comparison curve, namely the first texture curve.

[0078] After constructing the two curves, the Fletcher distance algorithm is used to calculate the structural matching between the two curves. Using the trace image texture curve and the first tool texture curve as input, the maximum offset distance between them during the traversal process is calculated. This distance value, as a quantitative expression of the structural difference, accurately reflects the overall matching degree of the two texture trends. This distance value is then normalized, using maximum distance normalization to convert it to the standard range of [0, 1]. This results in the final first texture similarity, which represents the degree of structural matching between the texture feature value set and the first texture feature value set in the texture dimension.

[0079] Step S700: When the first texture similarity reaches a predetermined similarity limit, the first type corresponding to the first tool is used as the tool type of the trace image.

[0080] In this embodiment of the present application, the first texture similarity calculated in the previous step is determined to determine whether it meets a predetermined similarity limit, thereby determining whether the current trace image can be determined to have been formed by a specific tool. The first tool is a candidate tool sample randomly selected from a tool feature database or automatically selected according to the recognition process sequence, and does not represent a unique tool. The first type is the tool classification label corresponding to the tool in the database, which is used to describe its structural category or usage attributes.

[0081] The recognition process first calculates the texture similarity between the current trace image and the first tool, and compares the result with a pre-set similarity threshold. This threshold is set by technical experts to distinguish between matching and mismatching texture structures, for example, at 0.8, to ensure sufficient recognition confidence.

[0082] When the similarity value is greater than or equal to this threshold, the texture features of the current trace image are determined to have a high degree of structural consistency with the standard template of the first tool, indicating that the tool is likely the source of the trace. At this point, the tool type corresponding to the first tool, that is, its classification label in the database, is used as the recognition result of the trace image, thereby determining the tool type corresponding to the image. This process automatically matches the trace image with the tool type through threshold control logic, automatically completing the intelligent classification and attribution of tool traces.

[0083] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0084] The present application extracts color features from a collected trace image to obtain a color feature value, wherein the trace image refers to an image of a tool's use trace on a trace-bearing object; matches the reference color feature value set of the trace-bearing object, and compares the color feature value with the reference color feature value set to obtain a color similarity set; determines whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; if so, issues a texture feature collection instruction, and obtains a texture feature value set of the trace image according to the texture feature collection instruction; extracts a first texture feature value set corresponding to a first tool in a tool feature database; compares the texture feature value set with the first texture feature value set to obtain a first texture similarity; when the first texture similarity reaches a predetermined similarity limit, the first type corresponding to the first tool is used as the tool type of the trace image. The present invention solves the technical problem in the prior art that the trace image is interfered with by factors such as material, lighting, corrosion, or thermal discoloration, resulting in poor recognition accuracy and inability to stably determine the tool type. By introducing a dual feature fusion and discrimination mechanism of color and texture, the trace image is dynamically matched with the tool template feature, achieving the technical effect of improving the accuracy and stability of tool trace recognition under complex conditions.

[0085] Embodiment 2 is based on the same inventive concept as the dynamic matching and recognition method of multimodal trace morphological features in the above embodiment. Figure 2 As shown, the present application provides a dynamic matching and recognition system for multimodal trace morphological features. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0086] The feature extraction module 11 is configured to extract color features from the collected trace image to obtain color feature values, wherein the trace image refers to an image of a use trace of a tool on a trace-bearing object; the matching and comparison module 12 is configured to match a reference color feature value set of the trace-bearing object, and compare the color feature values with the reference color feature value set to obtain a color similarity set; the judgment module 13 is configured to judge whether a color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; the texture feature collection module 14 is configured to issue a texture feature collection instruction if the color coefficient meets the predetermined coefficient threshold, and obtain a texture feature value set of the trace image according to the texture feature collection instruction; the extraction module 15 is configured to extract a first texture feature value set corresponding to a first tool in a tool feature database; the similarity determination module 16 is configured to compare the texture feature value set with the first texture feature value set to obtain a first texture similarity; and the tool type confirmation module 17 is configured to confirm a first type corresponding to the first tool as a tool type of the trace image when the first texture similarity meets a predetermined similarity limit value.

[0087] Further, the system is further configured to implement the following functions:

[0088] The system is further configured to implement the following functions:

[0089] Further, the system is further configured to implement the following functions:

[0090] The system is further configured to implement the following functions:

[0091] Further, the system is further configured to implement the following functions:

[0092] extracting a first color similarity in the color similarity set, wherein the first color similarity refers to a similarity of the color feature value and the original color feature value; extracting a second color similarity in the color similarity set, wherein the second color similarity refers to a color similarity of the trace image and a target eroded image in the eroded image set, and the target eroded image refers to an eroded image with the highest similarity to the color feature value; extracting a third color similarity in the color similarity set, wherein the third color similarity refers to a color similarity of the trace image and a target high-temperature image in the high-temperature image set, and the target high-temperature image refers to a high-temperature image with the highest similarity to the color feature value; obtaining the color coefficient according to the first color similarity, the second color similarity and the third color similarity.

[0093] Further, the system is further configured to implement the following functions:

[0094] obtaining a mean value of the second color similarity and the third color similarity, denoted as a first mean value; taking a mean value of the first mean value and the first color similarity, and performing normalization processing to obtain the color coefficient.

[0095] Further, the system is further configured to implement the following functions:

[0096] performing rasterization processing on the trace image to obtain a set of raster blocks; performing uniform sampling on the set of raster blocks based on a predetermined sampling strategy to obtain a set of target sample blocks; extracting a first sample block in the set of target sample blocks, and taking a first alternating current coefficient in a first discrete cosine transform coefficient of the first sample block as a first texture feature value; and assembling the set of texture feature values based on the first texture feature value according to the texture feature collection instruction.

[0097] Further, the system is further configured to implement the following functions:

[0098] performing curve processing on the set of texture feature values and the first set of texture feature values in sequence to obtain a texture curve and a first texture curve, respectively; and taking a first Fréchet distance of the texture curve and the first texture curve to represent the first texture similarity.

[0099] Further, the system is further configured to implement the following functions:

[0100] if not, issuing a repair instruction; and performing repair processing on the trace image based on the repair instruction and an image repair plan; wherein the image repair plan includes a physical repair plan and / or a chemical repair plan, and the physical repair plan includes plasma cleaning and laser cleaning, and the chemical repair plan includes composite pickling and oxidation-reduction.

[0101] Further, the system is also used to realize the following functions:

[0102] reading predetermined object features, and collecting multi-dimensional features of the object based on the predetermined object features to obtain an object feature set; performing normalized weighted analysis on the object feature set to obtain an object performance feedback coefficient; and adjusting the color coefficient by using the object performance feedback coefficient as a weight.

[0103] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above-mentioned specific embodiments of the present application have been described. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0104] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0105] The present specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that fall within the scope of the present application should be considered. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A dynamic matching and recognition method for multimodal trace morphological features, characterized in that: include: Performing color feature extraction on the collected trace image to obtain a color feature value, wherein the trace image refers to an image of the trace of use of the tool on the trace-bearing object; Matching a reference color feature value set of the mark-bearing object, and comparing the color feature value with the reference color feature value set to obtain a color similarity set; Determining whether a color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; If it is consistent, issuing a texture feature collection instruction, and obtaining a texture feature value set of the trace image according to the texture feature collection instruction; Extracting a first texture feature value set corresponding to a first tool from a tool feature database; Comparing the texture feature value set with the first texture feature value set to obtain a first texture similarity; When the first texture similarity reaches a predetermined similarity limit, taking the first type corresponding to the first tool as the tool type of the trace image; The reference color feature value set matching the trace-bearing object includes: Obtaining an original image of the mark-bearing object and obtaining an original color feature value of the original image; assembling a corrosion image set of the scar-bearing object, and sequentially obtaining a corrosion color feature value set of each corrosion image in the corrosion image set; assembling a high-temperature image set of the scar-bearing object, and sequentially obtaining a high-temperature color feature value set of each high-temperature image in the high-temperature image set; Building the reference color feature value set based on the primary color feature value, the corrosion color feature value set, and the high temperature color feature value set; Before determining whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold, the method includes: Extracting a first color similarity from the color similarity set, wherein the first color similarity refers to a similarity between the color feature value and the primary color feature value; Extracting a second color similarity from the color similarity set, wherein the second color similarity refers to a color similarity between the trace image and a target corrosion image in the corrosion image set, and the target corrosion image refers to a corrosion image having the highest similarity to the color feature value; Extracting a third color similarity from the color similarity set, wherein the third color similarity refers to a color similarity between the trace image and a target high-temperature image in the high-temperature image set, and the target high-temperature image refers to a high-temperature image having the highest similarity to the color feature value; Obtaining the color coefficient according to the first color similarity, the second color similarity, and the third color similarity; The step of obtaining the color coefficient according to the first color similarity, the second color similarity, and the third color similarity includes: Obtaining an average of the second color similarity and the third color similarity, recorded as a first average; The first mean value and the mean value of the first color similarity are obtained, and normalized to obtain the color coefficient.

2. The method according to claim 1, characterized in that Perform color feature extraction on the collected trace image to obtain color feature values, including: Obtaining discrete cosine transform coefficients of the trace image; The color feature value is represented by a DC coefficient in the discrete cosine transform coefficient.

3. The method according to claim 1, characterized in that Acquiring a texture feature value set of the trace image according to the texture feature collection instruction includes: Performing rasterization processing on the trace image to obtain a raster block set; Uniformly sampling the grid block set based on a predetermined sampling strategy to obtain a target sample block set; Extracting a first sample block from the target sample block set, and using a first AC coefficient in a first discrete cosine transform coefficient of the first sample block as a first texture feature value; The texture feature value set is constructed based on the first texture feature value according to the texture feature collection instruction.

4. The method according to claim 1, characterized in that Comparing the texture feature value set with the first texture feature value set to obtain a first texture similarity includes: performing curve processing on the texture feature value set and the first texture feature value set in sequence to obtain a texture curve and a first texture curve respectively; The first texture similarity is represented by a first Fletcher distance between the texture curve and the first texture curve.

5. The method according to claim 1, characterized in that: After determining whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold, the method further includes: If it does not meet the requirements, issue a repair instruction; Retrieving an image restoration plan based on the restoration instruction to restore the trace image; The image restoration plan includes a physical restoration plan and / or a chemical restoration plan, and the physical restoration plan includes plasma cleaning and laser cleaning, and the chemical restoration plan includes composite acid cleaning and oxidation-reduction.

6. The method according to claim 1, characterized in that Before determining whether the color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold, the method further includes: Reading predetermined object features, and collecting multi-dimensional features of the scar-bearing object based on the predetermined object features to obtain an object feature set; Performing normalized weighted analysis on the object feature set to obtain an object performance feedback coefficient; The color coefficient is adjusted using the object performance feedback coefficient as a weight.

7. A dynamic matching and recognition system for multimodal trace morphological features, characterized by: A system for performing a dynamic matching and recognition method for multimodal trace morphological features according to any one of claims 1 to 6, comprising: a feature extraction module, configured to extract color features from the collected trace image to obtain a color feature value, wherein the trace image is an image of the trace left by the tool on the trace-bearing object; a matching and comparison module, configured to match a reference color feature value set of the mark-bearing object and compare the color feature value with the reference color feature value set to obtain a color similarity set; A judgment module, configured to judge whether a color coefficient obtained by analyzing the color similarity set meets a predetermined coefficient threshold; A texture feature collection module, configured to issue a texture feature collection instruction if the condition is met, and obtain a texture feature value set of the trace image according to the texture feature collection instruction; An extraction module, configured to extract a first texture feature value set corresponding to a first tool in a tool feature database; a similarity determination module, configured to compare the texture feature value set with the first texture feature value set to obtain a first texture similarity; The tool type confirmation module is configured to use the first type corresponding to the first tool as the tool type of the trace image when the first texture similarity reaches a predetermined similarity limit.

Images