A shape reconstruction method combining low-frequency shapes and high-frequency normal vectors

Abstract

The present invention relates to the technical field of machine wear state monitoring, and in particular to a morphology reconstruction method combining low-frequency shape and high-frequency normal vectors. The method comprises the following steps: obtaining the clearest imaging frame number and corresponding height of the wear surface point through focus measurement and frame displacement estimation, generating low-frequency shape features of the wear surface area to be measured; establishing a normal vector calculation model for the wear surface with a large height difference based on a photometric image sequence of different focus areas, evaluating the focus area to aggregate multiple normal vectors, and achieving a clear description of the wear characteristics of the entire surface; and achieving three-dimensional morphology reconstruction of the wear surface based on the low-frequency shape features and the aggregated normal vectors. The present invention introduces low-frequency shape features to combat the warping deformation of the reconstructed morphology, and fuses the normal vectors of different focus areas to achieve an accurate description of the fine features of the surface with a large height difference, thereby improving the reconstruction accuracy of the microscopic morphology of the wear surface.

Description

Technical Field

[0001] The present invention relates to the technical field of machine wear state monitoring, and in particular to a morphology reconstruction method combining low-frequency shapes and high-frequency normal vectors. Background Art

[0002] Accumulated wear on friction pairs during the operation of mechanical equipment can limit component performance or even lead to failure, becoming a major factor restricting the equipment's operational life. As direct evidence of wear, the complex morphology of worn surfaces, which record key information such as the wear mechanism and extent, is crucial for characterizing the wear state of key components such as bearings and gears. Therefore, developing wear surface analysis methods is of great engineering significance for monitoring the operating status and diagnosing faults in mechanical equipment.

[0003] Existing wear surface analysis methods typically use instruments such as industrial endoscopes and optical microscopes to obtain two-dimensional images of the worn surface and empirically assess the wear surface condition. However, due to the lack of height information in two-dimensional images, these analysis methods cannot fully characterize the topography of the worn surface, making it impossible to quantify the severity of wear. While stereo imaging devices such as laser confocal microscopes and surface profilometers can provide comprehensive three-dimensional surface topography for wear surface analysis, they require component disassembly and segmentation, making them unsuitable for on-machine inspection and limiting their effectiveness in practical engineering applications. Therefore, current wear surface analysis incorporates computer vision methods to achieve in-situ three-dimensional topography measurement of the worn surface. Compared to binocular vision and structured light methods, photometric stereo vision can achieve pixel-level topography measurement using only a monocular microscope and multiple point light sources, providing rich topographic features for wear surface analysis. However, the limited depth of field of the microscope blurs high-frequency details of the local topography, and the accumulation of gradient noise during the reconstruction process induces low-frequency warping of the reconstructed surface, compromising the accurate description of the wear surface topography.

[0004] In summary, the application of photometric stereo vision extends the 3D topography measurement of worn surfaces to in-situ measurement scenarios, providing a comprehensive description of the topography for wear surface analysis. However, the limited imaging depth of field and the accumulation of gradient noise lead to high-frequency blurring and low-frequency distortion in the topography reconstruction results. Therefore, it is urgent to optimize photometric stereo reconstruction methods to provide accurate topography for wear severity assessment. Summary of the Invention

[0005] In order to solve the technical problems existing in the above-mentioned prior art, the present invention provides a morphology reconstruction method that combines low-frequency shapes and high-frequency normal vectors to solve the problems of high-frequency blur and low-frequency distortion in the existing photometric stereo reconstruction in wear surface analysis, and realize the accurate reconstruction of the three-dimensional morphology of the wear surface.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] In a first aspect, in one embodiment provided by the present invention, a method for reconstructing a shape by combining low-frequency shapes and high-frequency normal vectors is provided, the method comprising the following steps:

[0008] The clearest imaging frames and corresponding heights of the worn surface points are obtained through focus measurement and frame displacement estimation, and low-frequency shape features of the wear surface area to be measured are generated.

[0009] Based on the photometric image sequences of different focus areas, a normal vector calculation model for the wear surface with large height drop is established. The focus area is evaluated to aggregate multiple normal vectors to achieve a clear description of the wear characteristics of the entire surface.

[0010] Based on the low-frequency shape features and the aggregated normal vectors, three-dimensional topography reconstruction of the worn surface is achieved.

[0011] As a further solution of the present invention, before the step of obtaining the clearest imaging frames and corresponding heights of the wear surface points by focusing measurement and frame displacement estimation, and generating low-frequency shape features of the wear surface area to be measured, the step further includes:

[0012] Relying on the multi-focus image sequence in the wear surface imaging process, a wear surface low-frequency shape feature extraction model is established.

[0013] As a further solution of the present invention, the method of obtaining the clearest imaging frames and corresponding heights of the worn surface points by focusing measurement and frame displacement estimation to generate low-frequency shape features of the wear surface area to be measured includes the following steps:

[0014] The full-light source observation video collected during the focusing process of the wear surface microscopic imaging is decomposed frame by frame into a multi-focus image sequence;

[0015] Based on the focused image sequence, the clarity of the wear surface point features is described to obtain a wear image; a focus evaluation function is generated according to the wear image, and the clearest imaging frame number k0(x,y) is determined;

[0016] The displacement of the wear surface imaging system corresponding to each focused image sequence is solved from the imaging field of view, and the displacement z of the clearest imaging frame number image is determined. k ;

[0017] The low-frequency shape feature Z0 of the worn surface is obtained according to the clearest imaging frame number and the displacement of the frame image.

[0018] As a further solution of the present invention, the focus evaluation function is calculated by the following formula:

[0019]

[0020] Among them, M k (x, y) is the quantitative value of the focus degree of the k-th frame focused image at (x, y). The larger the value, the clearer the image at that pixel. For the high-frequency subband coefficients at level j and direction l after decomposition, the number of directions at each decomposition level is set to [1 41], that is, the total number of decompositions J = 3.

[0021] As a further solution of the present invention, the process of decomposing the full-light source observation video collected during the focusing process of the wear surface microscopic imaging into a multi-focus image sequence frame by frame also includes:

[0022] Taking the middle frame image of the multi-focus image sequence as the reference, each focused image is aligned with it through image phase registration;

[0023] Describe the image registration results.

[0024] As a further solution of the present invention, the step of describing the clarity of point features on the worn surface based on the focused image sequence to obtain a wear image; generating a focus evaluation function based on the wear image and determining the number of clearest imaging frames includes:

[0025] A focus evaluation function is constructed to quantify each focused image sequence to describe the clarity of the wear surface point features and obtain the wear image.

[0026] Decomposing the wear image into high-frequency sub-band and low-frequency sub-band images by using a non-subsampled shearing transform, and generating a focus evaluation function;

[0027] The image frame number corresponding to the maximum focus evaluation function value is selected pixel by pixel to determine the clearest imaging frame number.

[0028] As a further solution of the present invention, the number of the clearest imaging frames k0(x, y) and the displacement z of the frame image are k , obtain the low-frequency shape feature Z0 of the worn surface, including:

[0029] According to the clearest imaging frame number k0(x,y) and the displacement z of the frame image k , determine the surface height value of each point on the worn surface pixel by pixel;

[0030] The extreme value noise of the surface height value is removed by mean filtering, and the low-frequency shape feature Z0 of the worn surface is obtained.

[0031] As a further solution of the present invention, the three-dimensional morphology reconstruction of the worn surface based on the low-frequency shape features and the aggregated normal vectors includes:

[0032] The wear surface reconstruction cost function is designed by combining the low-frequency shape features and the aggregated normal vectors. The optimal height distribution is sought by minimizing the cost function to achieve three-dimensional morphology reconstruction of the wear surface.

[0033] As a further solution of the present invention, the method of designing a wear surface reconstruction cost function by combining the low-frequency shape features and the aggregated normal vectors, and seeking an optimal height distribution by minimizing the cost function to achieve three-dimensional morphology reconstruction of the wear surface includes:

[0034] The high-frequency normal vector is converted into the wear surface gradient {p(x,y),q(x,y)}, and a joint cost function is constructed to quantitatively describe the difference between the known surface gradient, low-frequency shape and the height distribution of the optimized surface to obtain the difference result;

[0035] Based on the difference results, the conjugate gradient method is used to minimize the nonlinear cost function to solve the optimal height distribution that satisfies the low-frequency shape and high-frequency normal vector, and obtain the accurate three-dimensional morphology of the worn surface.

[0036] As a further solution of the present invention, the joint cost function is calculated by the following formula to obtain the difference result:

[0037]

[0038] Where J(Z) is the quantitative evaluation result of the difference between the optimized surface Z and the calculated low-frequency shape Z0 and high-frequency gradient (p, q); Ω is the area to be reconstructed; Z x ,Z y To optimize the partial derivative of surface Z in the (x, y) direction, that is Z y =D y Z and D are numerical differential matrices; λ is the regularization coefficient, whose value is determined by the U-curve method; G(Z) is the low-frequency characteristic component of the optimized surface extracted by Gaussian blur processing.

[0039] Compared with the prior art, the present invention has at least the following beneficial effects:

[0040] (1) This paper extracts low-frequency shape features of worn surfaces from multi-focus image sequences through focus measurement and frame displacement estimation. This approach eliminates the need for additional sensors or precision stepper motors and can be applied to in-situ photometric stereo reconstruction scenarios. Furthermore, the extracted low-frequency shape maintains the same resolution as the normal vector, simplifying subsequent joint reconstruction.

[0041] (2) The present invention aggregates the surface normal vector features of different focus areas through normal vector focusing measurement and fusion, thereby achieving an accurate description of the direction of micro-elements on surfaces with large height differences, solving the problem of high-frequency normal vector blurring caused by the limited imaging depth of field of digital microscopes, and realizing the description of three-dimensional morphology.

[0042] These and other aspects of the present invention will become more readily apparent in the following description of the embodiments. It should be understood that the above general description and the following detailed description are merely exemplary and explanatory and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the flow of the morphology reconstruction method that combines low-frequency shapes and high-frequency normal vectors.

[0044] Figure 2 Schematic diagram of the structure of step S10 in the method for reconstructing the morphology by combining low-frequency shapes and high-frequency normal vectors.

[0045] Figure 3 Schematic diagram of the structure of step S20 in the method for reconstructing the morphology by combining low-frequency shapes and high-frequency normal vectors.

[0046] Figure 4 Schematic diagram of the structure of step S30 in the method for reconstructing the morphology by combining low-frequency shapes and high-frequency normal vectors.

[0047] Figure 5 Schematic diagram of the specific process of the morphology reconstruction method combining low-frequency shapes and high-frequency normal vectors.

[0048] Figure 6 It is a multi-focus image sequence of the wear surface area to be measured.

[0049] Figure 7 This is an imaging model for multifocus images, which describes the relationship between the imaging field of view and the displacement of the imaging system.

[0050] Figure 8 It is a pseudo height map that describes the distribution of the maximum number of clear imaging frames at different pixel points.

[0051] Figure 9 This is the final low-frequency shape map that describes the overall morphology of the worn surface.

[0052] Figure 10 is the calculation result of the normal vector corresponding to the remote imaging. When the imaging camera is far away from the surface to be measured, the photometric image sequence achieves clear imaging on the raised part of the worn surface. The normal vector calculated based on the photometric image sequence has rich detail features in the corresponding focus area.

[0053] Figure 11 The normal vector calculation result corresponding to the near-end imaging is calculated from the photometric image sequence collected when the imaging camera is close to the surface to be measured.

[0054] Figure 12 The fused normal vector can accurately describe the subtle morphological features of the entire area.

[0055] Figure 13 This is the topography reconstruction result of the measured area on the worn surface. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0058] It should be understood that the terms used in this specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the specification and appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0059] Specifically, the embodiments of the present invention are further described below with reference to the accompanying drawings.

[0060] See also Figure 1 and Figure 5 , Figure 1 Flowchart of a method for morphology reconstruction combining low-frequency shape and high-frequency normal vector provided by an embodiment of the present invention. Figure 1 As shown, the method for reconstructing the shape by combining low-frequency shapes and high-frequency normal vectors includes steps S10 to S30.

[0061] S10. Obtain the clearest imaging frame number and corresponding height of the wear surface point through focus measurement and frame displacement estimation, and generate low-frequency shape features of the wear surface area to be measured.

[0062] Before the step of obtaining the clearest imaging frames and corresponding heights of the wear surface points by focusing measurement and frame displacement estimation, and generating low-frequency shape features of the wear surface area to be measured, the following steps are further included:

[0063] Relying on the multi-focus image sequence in the wear surface imaging process, a wear surface low-frequency shape feature extraction model is established.

[0064] See also Figure 2In an embodiment of the present invention, the steps of obtaining the clearest imaging frames and corresponding heights of the wear surface points by focusing measurement and frame displacement estimation to generate low-frequency shape features of the wear surface area to be measured include the following steps:

[0065] See also Figure 6-9 , S101, decompose the full light source observation video collected during the focusing process of the wear surface microscopic imaging into a multi-focus image sequence frame by frame.

[0066] In an embodiment of the present invention, the step of decomposing the full-light source observation video collected during the focusing process of the wear surface microscopic imaging into a multi-focus image sequence frame by frame further includes:

[0067] Using the intermediate frame of a multifocus image sequence as a reference, each focused image is aligned with it through image phase registration. Specifically, the image registration results are described. Taking into account the plane position movement and working distance changes of the digital microscope, the image registration results can be described using affine transformation matrices for translation and scaling.

[0068] S102 , based on the focused image sequence, describe the clarity of the wear surface point features to obtain a wear image; generate a focus evaluation function according to the wear image, and determine the clearest imaging frame number k0(x,y).

[0069] In an embodiment of the present invention, the steps of describing the clarity of point features on the worn surface based on the focused image sequence to obtain a wear image, generating a focus evaluation function based on the wear image, and determining the clearest imaging frame number k0(x,y) include:

[0070] S1021, constructing a focus evaluation function to quantify each focused image sequence to describe the clarity of the wear surface point features, thereby obtaining a wear image;

[0071] The focusing evaluation function is calculated by the following formula:

[0072]

[0073] Among them, M k (x, y) is the quantitative value of the focus degree of the k-th frame focused image at (x, y). The larger the value, the clearer the image at that pixel. The high-frequency subband coefficients at level j and direction l after decomposition are shown in Table 1. The number of directions at each decomposition level is set to [1 4 1], that is, the total number of decompositions is J = 3. L represents the number of decomposition directions at the next highest level under the current decomposition condition.

[0074] S1022, decomposing the wear image into high-frequency sub-band and low-frequency sub-band images using a non-subsampling shearing transform, and generating a focus evaluation function;

[0075] S1023 , selecting the image frame number corresponding to the maximum focus evaluation function value pixel by pixel, and determining the clearest imaging frame number k0 (x, y).

[0076] S103, solving the corresponding wear surface imaging system displacement from the imaging field of each focused image sequence, and determining the displacement z of the clearest imaging frame image k .

[0077] Specifically, the relative size of the wear surface imaging field of view is quantitatively characterized according to the scaling factor in the image registration, and then the position change of the wear surface imaging system is determined according to the pinhole model, as follows:

[0078]

[0079] Among them, s k is the zoom factor of the k-th frame focused image compared to the reference image; k is the displacement of the surface to be measured compared to the optical center plane during the kth imaging; n is the ratio of the sensor size of the digital microscope to the imaging focal length.

[0080] S104, according to the clearest imaging frame number k0 (x, y) and the displacement z of the frame image k , obtain the low-frequency shape feature Z0 of the worn surface.

[0081] In the embodiment of the present invention, the number of the clearest imaging frames k0(x, y) and the displacement z of the frame image are k , obtain the low-frequency shape feature Z0 of the worn surface, including:

[0082] S1041, based on the clearest imaging frame number k0 (x, y) and the displacement z of the frame image k , determine the surface height value of each point on the worn surface pixel by pixel;

[0083] S1042: Remove extreme noise of the surface height value by mean filtering, and obtain a low-frequency shape feature Z0 of the worn surface.

[0084] This method uses focus metrics and frame displacement estimation to extract low-frequency shape features of worn surfaces from multi-focus image sequences. This approach eliminates the need for additional sensors or precision stepper motors and can be applied to in-situ photometric stereo reconstruction scenarios. Furthermore, the extracted low-frequency shape maintains the same resolution as the normal vector, simplifying subsequent joint reconstruction.

[0085] See also Figure 10-12 ,S20, based on the photometric image sequences of different focus areas, a normal vector calculation model for the wear surface with large height difference is established, and the focus area is evaluated to aggregate multiple normal vectors to achieve a clear description of the wear characteristics of the entire surface.

[0086] See also Figure 3 In an embodiment of the present invention, the normal vector calculation model of a large height drop wear surface is established based on a sequence of photometric images of different focus areas, and the focus area is evaluated to aggregate multiple normal vectors to achieve a clear description of the wear characteristics of the entire surface, including:

[0087] S201 , adjusting the imaging distance of the imaging system multiple times, capturing photometric image sequences of the worn surface in different focus areas, and calculating corresponding normal vectors.

[0088] S202: Construct a focus evaluation function for three-dimensional vectors to obtain the clarity of each normal vector. The focus evaluation function is calculated using the following formula:

[0089]

[0090] in, is the quantitative value of the focusing degree of the normal vector at (x, y) calculated for the kth group of photometric image sequences, Represents the normal vector components [N x (x,y)N y (x,y)N z (x,y)] is the focus evaluation function value at (x,y).

[0091] S203. Based on the clarity of each normal vector and according to the maximum principle of the focusing evaluation function, the normal vector of each wear feature is selected to achieve a clear description of the high-frequency features of the wear surface with a large height difference over the entire domain.

[0092] The present invention aggregates the surface normal vector features of different focus areas through normal vector focusing measurement and fusion, achieves accurate description of the direction of micro-elements on surfaces with large height differences, and solves the problem of high-frequency normal vector blurring caused by the limited imaging depth of field of digital microscopes.

[0093] See also Figure 13 , S30, based on the low-frequency shape features and the aggregated normal vector, realize three-dimensional morphology reconstruction of the worn surface.

[0094] The three-dimensional morphology reconstruction of the worn surface based on the low-frequency shape features and the aggregated normal vectors includes:

[0095] The wear surface reconstruction cost function is designed by combining the low-frequency shape features and the aggregated normal vectors. The optimal height distribution is sought by minimizing the cost function to achieve three-dimensional morphology reconstruction of the wear surface.

[0096] See also Figure 4In an embodiment of the present invention, the design of a wear surface reconstruction cost function by combining the low-frequency shape features and the aggregated normal vectors, and the search for an optimal height distribution by minimizing the cost function to achieve three-dimensional topography reconstruction of the wear surface include:

[0097] S301. Convert the high-frequency normal vector into the wear surface gradient {p(x,y),q(x,y)}, construct a joint cost function, quantify the difference between the known surface gradient, low-frequency shape and the height distribution of the optimized surface, and obtain the difference result.

[0098] Among them, the joint cost function is calculated by the following formula to obtain the difference result:

[0099]

[0100] Where J(Z) is the quantitative evaluation result of the difference between the optimized surface Z and the calculated low-frequency shape Z0 and high-frequency gradient (p, q); Ω is the area to be reconstructed; Z x ,Z y To optimize the partial derivative of surface Z in the (x, y) direction, that is Z y =D y Z and D are numerical differential matrices; λ is the regularization coefficient, whose value is determined by the U-curve method; G(Z) is the low-frequency characteristic component of the optimized surface extracted by Gaussian blur processing.

[0101] S302 : Based on the difference results, a conjugate gradient method is used to minimize a nonlinear cost function, and an optimal height distribution that satisfies the low-frequency shape and high-frequency normal vector is solved to obtain an accurate three-dimensional morphology of the worn surface.

[0102] The present invention obtains a multi-focus image sequence during the focusing process and extracts low-frequency features of the worn surface from the multi-focus image sequence through focus measurement and frame displacement estimation. A photometric image sequence of different focus areas is captured, and then high-frequency normal vectors of the surface are calculated and aggregated to accurately describe the detailed features of the worn surface. The normal vectors are converted into surface gradients, and a cost function combining high- and low-frequency features is constructed. The optimal surface height distribution is selected through optimization to accurately reconstruct the three-dimensional morphology of the worn surface. The present invention introduces low-frequency shape features to combat warping deformation of the reconstructed morphology and integrates normal vectors from different focus areas to accurately describe the fine features of surfaces with large height differences, thereby improving the reconstruction accuracy of the microscopic morphology of the worn surface.

[0103] It should be understood that, although the above is described in a certain order, these steps are not necessarily performed in sequence according to the above order. Unless there is clear explanation in this article, the execution of these steps does not have strict order restriction, and these steps can be performed in other orders. Moreover, a part of the steps of the present embodiment may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be performed in turn or alternately with at least a portion of the steps or stages in other steps or other steps.

[0104] The above are exemplary embodiments disclosed in the present invention, but it should be noted that various changes and modifications may be made without departing from the scope of the embodiments disclosed in the claims. The functions, steps and / or actions of the method claims according to the disclosed embodiments described herein do not need to be performed in any particular order. In addition, although the elements disclosed in the embodiments of the present invention may be described or required in individual form, they may also be understood as multiple unless expressly limited to the singular.

[0105] It should be understood that, as used herein, the singular form "a" or "an" is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and / or" refers to any and all possible combinations of one or more of the items listed in association. The serial numbers of the embodiments disclosed in the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0106] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the disclosure of the embodiments of the present invention (including the claims) is limited to these examples. Within the spirit of the embodiments of the present invention, the technical features of the above embodiments or different embodiments may be combined, and there are many other variations of different aspects of the above embodiments of the present invention, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the scope of protection of the embodiments of the present invention.

Claims

1. A method for morphology reconstruction combining low-frequency shapes and high-frequency normal vectors, characterized in that: The method includes: The clearest imaging frames and corresponding heights of the worn surface points are obtained through focus measurement and frame displacement estimation, and low-frequency shape features of the wear surface area to be measured are generated. Based on the photometric image sequences of different focus areas, a normal vector calculation model for the wear surface with large height drop is established. The focus area is evaluated to aggregate multiple normal vectors to achieve a clear description of the wear characteristics of the entire surface. Reconstructing the three-dimensional morphology of the worn surface based on the low-frequency shape features and the aggregated normal vectors; The method of obtaining the clearest imaging frames and corresponding heights of the wear surface points by focusing measurement and frame displacement estimation, and generating low-frequency shape features of the wear surface area to be measured, includes the following steps: The full-light source observation video collected during the focusing process of the wear surface microscopic imaging is decomposed frame by frame into a multi-focus image sequence; Based on the focused image sequence, the clarity of the wear surface point features is described to obtain a wear image; a focus evaluation function is generated according to the wear image, and the clearest imaging frame number k0(x,y) is determined; The displacement of the wear surface imaging system corresponding to each focused image sequence is solved from the imaging field of view, and the displacement z of the clearest imaging frame number image is determined. k ; The low-frequency shape feature Z0 of the worn surface is obtained according to the clearest imaging frame number and the displacement of the frame image.

2. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, characterized in that: Before the step of obtaining the clearest imaging frames and corresponding heights of the wear surface points by focusing measurement and frame displacement estimation and generating low-frequency shape features of the wear surface area to be measured, the step further includes: Relying on the multi-focus image sequence in the wear surface imaging process, a wear surface low-frequency shape feature extraction model is established.

3. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, characterized in that: The focusing evaluation function is calculated by the following formula: Among them, M k (x, y) is the quantitative value of the focus degree of the k-th frame focused image at (x, y). The larger the value, the clearer the image at that pixel. For the high-frequency subband coefficients at level j and direction l after decomposition, the number of directions at each decomposition level is set to [1 4 1], that is, the total number of decompositions J = 3.

4. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, characterized in that: The method of decomposing the full-light source observation video collected during the focusing process of the wear surface microscopic imaging into a multi-focus image sequence frame by frame also includes: Taking the middle frame image of the multi-focus image sequence as the reference, each focused image is aligned with it through image phase registration; Describe the image registration results.

5. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, characterized in that: The description of the clarity of point features on the worn surface based on the focused image sequence is used to obtain a wear image; Generating a focus evaluation function according to the wear image and determining the clearest imaging frame number, including: A focus evaluation function is constructed to quantify each focused image sequence to describe the clarity of the wear surface point features and obtain the wear image. Decomposing the wear image into high-frequency sub-band and low-frequency sub-band images by using a non-subsampled shearing transform, and generating a focus evaluation function; The image frame number corresponding to the maximum focus evaluation function value is selected pixel by pixel to determine the clearest imaging frame number.

6. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, wherein: The step of obtaining the low-frequency shape feature Z0 of the worn surface according to the clearest imaging frame number and the displacement of the frame image includes: According to the clearest imaging frame number k0(x,y) and the displacement z of the frame image k , determine the surface height value of each point on the worn surface pixel by pixel; The extreme value noise of the surface height value is removed by mean filtering, and the low-frequency shape feature Z0 of the worn surface is obtained.

7. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 1, characterized in that: The three-dimensional morphology reconstruction of the worn surface based on the low-frequency shape features and the aggregated normal vectors includes: The wear surface reconstruction cost function is designed by combining the low-frequency shape features and the aggregated normal vectors. The optimal height distribution is sought by minimizing the cost function to achieve three-dimensional morphology reconstruction of the wear surface.

8. The method for morphology reconstruction combining low-frequency shape and high-frequency normal vector according to claim 7, characterized in that: The method of designing a wear surface reconstruction cost function by combining the low-frequency shape features and the aggregated normal vectors, and seeking an optimal height distribution by minimizing the cost function to achieve three-dimensional morphology reconstruction of the wear surface includes: The high-frequency normal vector is converted into the wear surface gradient {p(x,y),q(x,y)}, and a joint cost function is constructed to quantitatively describe the difference between the known surface gradient, low-frequency shape and the height distribution of the optimized surface to obtain the difference result; Based on the difference results, the conjugate gradient method is used to minimize the nonlinear cost function to solve the optimal height distribution that satisfies the low-frequency shape and high-frequency normal vector, and obtain the accurate three-dimensional morphology of the worn surface.

9. The method for joint low-frequency shape and high-frequency normal vector morphology reconstruction according to claim 8, characterized in that: The joint cost function is calculated by the following formula to obtain the difference result: Where J(Z) is the quantitative evaluation result of the difference between the optimized surface Z and the calculated low-frequency shape Z0 and high-frequency gradient (p, q); Ω is the area to be reconstructed; Z x ,Z y To optimize the partial derivative of surface Z in the (x, y) direction, that is Z y =D y Z and D are numerical differential matrices; λ is the regularization coefficient, whose value is determined by the U-curve method; G(Z) is the low-frequency characteristic component of the optimized surface extracted by Gaussian blur processing.

Images