Titanium alloy finished ingot surface polishing quality evaluation method and system

Abstract

This invention relates to the field of image data processing technology, specifically to a method and system for evaluating the surface polishing quality of titanium alloy ingots. The method includes: acquiring original images of the titanium ingot from different viewpoints and polishing cycles; performing frequency domain transformation on the original images to obtain dark areas based on the spatial distribution of high-frequency components in the frequency domain; calculating the polishing removal degree based on the area of ​​the dark areas; obtaining the process completion degree using the polishing removal degree and the corresponding high-frequency components of the original images; determining the process switching weight for the current polishing cycle based on the process completion degree; and determining whether a polishing process switch is necessary based on the process switching weight for the current polishing cycle. This invention effectively overcomes the shortcomings of existing technologies, such as difficulties in cross-cycle registration, inaccurate boundary judgment, and reliance on a fixed number of process switches, significantly improving the surface quality consistency of titanium alloy ingots while reducing excessive consumption of expensive titanium materials.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and specifically to a method and system for evaluating the surface polishing quality of titanium alloy finished ingots. Background Technology

[0002] Titanium alloy ingots are key basic materials in aerospace and high-end manufacturing fields. When they are produced from the furnace, a rough, oxygen-rich, and highly hard oxide scale (the α-layer) naturally forms on their surface. Due to the high hardness and poor plasticity of the α-layer, it is prone to cracking under external forces, and these cracks can easily propagate into the material, potentially causing overall structural damage to the titanium alloy ingot. Therefore, before leaving the factory, titanium alloy ingots must undergo surface polishing to completely remove the oxide layer and eliminate any microcracks or other defects. Furthermore, given the high cost of titanium alloy materials, it is necessary to conduct real-time evaluation of the polishing quality to precisely control the polishing process, ensuring that the ingot meets finished product standards while minimizing material loss.

[0003] Currently, the industry generally adopts machine vision-based polishing quality assessment methods. A typical workflow involves deploying industrial cameras on the polishing production line to capture single-frame images of the titanium ingot surface after each polishing process (such as rough polishing, medium polishing, and fine polishing). Then, image processing algorithms (such as grayscale histogram analysis, texture feature extraction, and edge detection) are used to extract surface state features from the current image and compare them with a preset quality threshold. Based on this comparison, it is determined whether the current process is complete and whether it is permissible to proceed to the next process.

[0004] However, the aforementioned existing technologies have significant shortcomings in practical applications. First, during the polishing process, the surface polishing marks produced by different polishing passes change significantly (for example, coarse polishing uses coarse-grained abrasive belts, leaving deep and wide scratches; fine polishing uses fine-grained abrasive belts, resulting in shallower and denser scratches). Furthermore, the titanium ingot needs to rotate continuously during polishing to ensure uniform contact with the polishing tools, making it difficult to establish accurate pixel-level correspondences between images acquired from previous passes. Therefore, existing systems cannot reliably map the polished area in the current image back to the same physical area on the titanium ingot surface in previous passes, thus failing to perceive the continuous evolution of each local area as the polishing process progresses. Second, when polishing reaches the critical stage where the oxide layer is nearing complete removal, the same image will simultaneously contain the exposed titanium alloy body area (high reflectivity, bright field) and the residual oxide layer area (low reflectivity, dark field), with the two interspersed. At this point, methods relying on independent judgment of a single frame image cannot accurately quantify this mixed state, cannot distinguish between the residual oxide layer and the shadows caused by surface unevenness, and cannot determine whether continued polishing will lead to excessive wear of the titanium body. Due to the aforementioned shortcomings, the existing technology lacks sufficient accuracy in real-time progress perception at critical grinding positions, resulting in significant discrepancies between polishing quality assessment results and actual conditions. This ultimately affects the finished product qualification rate and causes unnecessary material waste. Summary of the Invention

[0005] This invention provides a method and system for evaluating the surface polishing quality of titanium alloy finished ingots, in order to solve existing problems.

[0006] The present invention provides a method and system for evaluating the surface polishing quality of titanium alloy finished ingots, which adopts the following technical solution: One embodiment of the present invention provides a method for evaluating the surface polishing quality of a finished titanium alloy ingot, the method comprising the following steps: Acquire original images of titanium ingots from different perspectives and under different polishing cycles; Frequency domain transformation is performed on the original images from different viewpoints and polishing cycles. Based on the spatial distribution of high-frequency components in the frequency domain, the dark field regions under different viewpoints and polishing cycles are obtained. Based on the area of ​​the dark field region under different viewing angles and different polishing rounds, the polishing removal rate under different viewing angles and different polishing rounds is calculated. By utilizing the polishing removal rate under different perspectives and polishing cycles, as well as the high-frequency components corresponding to the original images under different perspectives and polishing cycles, the process completion rate under different perspectives and polishing cycles can be obtained. The process switching weight for the current polishing round is determined based on the process completion rate under different perspectives and polishing rounds. Based on the process switching weight of the current polishing cycle, determine whether a polishing process switch is required.

[0007] Furthermore, the specific steps involved in performing frequency domain transformation on the original images from different viewpoints and polishing cycles, and obtaining the dark field regions from different viewpoints and polishing cycles based on the spatial distribution of high-frequency components in the frequency domain, are as follows: Regions of interest were extracted from the original images of titanium ingots under different viewpoints and polishing cycles to obtain region images; The region image is converted into a spectrum image using Fast Fourier Transform; Arrange the frequency components in the spectrum image in descending order of their amplitude values ​​to obtain the frequency-amplitude sequence. Dark field regions under different viewing angles and polishing cycles are obtained based on frequency values ​​in the frequency-amplitude sequence.

[0008] Furthermore, the specific steps for obtaining the dark field region under different viewing angles and polishing cycles based on the frequency values ​​in the frequency-amplitude sequence are as follows: Calculate the difference between the frequency values ​​of two adjacent frequency components in the frequency-amplitude sequence, and find the pair of frequency components with the largest absolute value of the difference; Take the larger frequency value in a pair of frequency components and the largest frequency value in the frequency-amplitude sequence as the interval endpoints, and extract all frequency components within the interval corresponding to the interval endpoints as the high-frequency part. The high-frequency component is subjected to inverse fast Fourier transform to obtain the reconstructed image; The regions with non-zero pixel values ​​in the reconstructed image are identified as polished regions, while the remaining pixel regions are identified as dark areas under different viewing angles and polishing cycles.

[0009] Furthermore, the specific steps for calculating the polishing removal rate based on the dark area area under different viewing angles and polishing cycles are as follows: The area of ​​the dark field region under the current view and the current polishing round is taken as the current dark field area, and the area of ​​the dark field region under the current view and the previous polishing round is taken as the previous dark field area. The area of ​​the dark field region at the beginning of the current polishing process is taken as the initial dark field area. The first ratio is obtained by dividing the difference between the previous dark field area and the current dark field area by the initial dark field area. The difference between 1 and the first ratio is taken as the polishing removal degree under the current view and the current polishing round; Obtain the polishing removal rate under different perspectives and different polishing cycles.

[0010] Furthermore, the specific steps for obtaining the process completion degree under different viewpoints and polishing cycles by utilizing the polishing removal degree under different viewpoints and polishing cycles, as well as the high-frequency components corresponding to the original images under different viewpoints and polishing cycles, are as follows: Add 1 to the polishing removal degree under the current view and the current polishing round to get the sum value, and subtract the polishing removal degree under the current view and the previous polishing round from the sum value to get the polishing removal degree improvement amount; For each frequency value in the high-frequency part under the current viewpoint and the current polishing round, calculate the absolute difference between this frequency value and each frequency value in the high-frequency part under the current viewpoint and the previous polishing round, and select the minimum absolute difference value. Calculate the mean frequency value of the high-frequency part under the current view and the current polishing round, and divide the minimum absolute difference by the mean frequency value to obtain the polishing intensity similarity of each frequency value in the high-frequency part under the current view and the current polishing round; The polishing intensity similarity of all frequency values ​​in the high-frequency part under the current view and the current polishing round is summed, and the sum is multiplied by the polishing removal improvement to obtain the process completion degree under the current view and the current polishing round. Obtain the process completion rate from different perspectives and with different polishing cycles.

[0011] Furthermore, the specific steps for determining the process switching weight of the current polishing round based on the process completion degree under different perspectives and polishing rounds are as follows: Obtain the process completion degree of all views under the current polishing wheel, and determine the minimum and average process completion degree. Divide the minimum process completion rate by the average process completion rate to obtain the degree of non-compliance of the process completion rate for the current polishing round; The first sum is obtained by squaring the difference between the process completion degree and the mean process completion degree of each viewpoint under the current polishing cycle; Divide the number of viewpoints by the first sum to get the quotient, and then take the square root of the quotient to get the overall deviation. Add 1 to the overall deviation to get the second sum. Multiply the degree of non-compliance of the current polishing cycle by the second sum to get the process switching weight of the current polishing cycle.

[0012] Furthermore, the specific steps for determining whether a polishing process switch is needed based on the process switching weight of the current polishing round are as follows: The process switching weight of the current polishing cycle is normalized. If the normalized value is greater than the preset switching threshold, the current polishing process ends and the process switches to the next polishing process; otherwise, the next polishing cycle of the current process continues.

[0013] Furthermore, the method also includes: The second ratio is obtained by counting the number of times the process completion rate is lowest from the current perspective in the current polishing process and dividing this number by the number of polishing rounds in the current polishing process. Obtain the maximum process completion rate of all views under the current polishing cycle, and the standard deviation of the process completion rate of the current view in the current polishing process; The difference is obtained by subtracting the process completion value of the current view and the current polishing cycle from the maximum process completion value of all views under the current polishing cycle. The difference is then divided by the standard deviation of the process completion value of the current view in the current polishing process to obtain the third ratio. The sum of the second and third ratios is used to determine the construction difficulty from the current perspective.

[0014] Furthermore, the method also includes: If the construction difficulty from the current perspective exceeds the first safety threshold, then the current perspective is determined to be a difficult perspective. If the construction difficulty of the current view is less than or equal to the first safety threshold and greater than the second safety threshold, then the current view is determined to be a view to be processed. If the construction difficulty of the current perspective is less than or equal to the second safety threshold, then the current perspective is determined to be a qualified perspective.

[0015] This invention proposes an evaluation system for the surface polishing quality of titanium alloy finished ingots, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the evaluation method for the surface polishing quality of titanium alloy finished ingots as described above.

[0016] The beneficial effects of the technical solution of this invention are as follows: This invention proposes a method and system for evaluating the surface polishing quality of titanium alloy ingots. By acquiring original images from different perspectives and polishing cycles and extracting dark areas based on frequency domain transformation, it achieves accurate quantitative analysis of the residual oxide layer state on the surface of titanium alloy ingots. The polishing removal degree is calculated based on the dark area, and the process completion degree is obtained by integrating the high-frequency component differences between adjacent cycles, thereby enabling the perception of the removal efficiency and texture convergence of the current process. Furthermore, based on the ratio of the minimum to the average process completion degree under all perspectives and the dispersion of completion degree at each perspective, a switching weight is adaptively determined and compared with a threshold after normalization to trigger process switching. These technical means enable the system to continuously track the polishing process of the same physical area, accurately separate the polishing influence zone and residual dark area at the critical oxide layer removal stage, and automatically determine the switching timing for rough polishing, medium polishing, and fine polishing based on actual processing capabilities. This effectively overcomes the defects of existing technologies, such as difficulty in cross-cycle registration, inaccurate boundary judgment, and reliance on a fixed number of process switching steps, significantly improving the surface quality consistency of titanium alloy ingots while reducing excessive consumption of expensive titanium materials. In addition, heat maps and quality reports are generated based on the construction difficulty, providing intuitive support for defect tracing and closed-loop control. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the steps of a method for evaluating the surface polishing quality of a titanium alloy finished ingot according to the present invention. Figure 2 This is a block diagram of an evaluation system for the surface polishing quality of a titanium alloy finished ingot according to the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method and system for evaluating the surface polishing quality of a titanium alloy ingot according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method and system for evaluating the surface polishing quality of titanium alloy finished ingots provided by this invention.

[0022] Please see Figure 1 The diagram illustrates a flowchart of a method for evaluating the surface polishing quality of a titanium alloy ingot according to an embodiment of the present invention. The method includes the following steps: Step S001: Obtain original images of the titanium ingot from different perspectives and under different polishing cycles.

[0023] It should be noted that titanium ingots refer to intermediate products that have not yet reached the quality standards of finished products. Their surfaces are covered with oxide scale (α-layer) and may contain microcracks or other defects, requiring grinding and polishing.

[0024] Titanium alloy finished ingot: refers to the final product that has completed all polishing processes, removed the surface oxide layer, and met the quality standards.

[0025] The deployment of the vision module and the polishing acquisition process should minimize modifications to the existing equipment and structure of the polishing production line. Therefore, in this embodiment, a high-definition industrial camera is deployed directly above the titanium ingot to acquire optical images of the ingot surface, obtaining an image before polishing.

[0026] Regarding the correspondence between the titanium ingot surface and the images: Since the titanium ingot rotates continuously during the grinding process, its spatial position changes constantly. Therefore, an angle sensor is used to acquire the rotation angle of the titanium ingot fixture in real time, and the shooting angle corresponding to each frame of the image is used as the marker for that image. Subsequently, based on the shooting sequence, images with the same angle marker are extracted, thus forming an image sequence of each surface of the titanium ingot during the processing.

[0027] In this embodiment, the titanium ingot to be polished is cylindrical, with a diameter of 300 mm and a height of 500 mm, and is clamped on a rotatable polishing fixture. A high-definition industrial camera (1920×1080 pixels, 30fps resolution) is fixedly installed 1.2 m directly above the titanium ingot, with the lens vertically downwards and aimed at the center of the top surface of the titanium ingot. A high-precision angle sensor (0.1° resolution) is installed on the fixture's rotating shaft to record the rotation angle of the titanium ingot in real time.

[0028] The system presets m fixed viewing angles, for example, m=8, with adjacent viewing angles spaced 45° apart (0°, 45°, 90°, 135°, 180°, 225°, 270°, 315°). Each viewing angle corresponds to an angle range (for example, ±22.5° centered on that angle). When the angle sensor reading falls into a certain range, the camera is triggered to capture an image.

[0029] Before each polishing process (such as rough polishing), the titanium ingot is first rotated sequentially to all preset viewing angles in a stationary manner, and a set of "baseline images" are acquired as an initial state reference for that process. The polishing process then begins: after each polishing cycle (i.e., one polishing round k), the system controls the titanium ingot to rotate sequentially to each preset viewing angle, and automatically captures an original image at each angle—that is, the original image of the titanium ingot under different viewing angles and polishing rounds. Simultaneously, the current polishing round number k (k=1,2,3,…) and the corresponding viewing angle number v (v=1,…,8) are recorded. All original images, along with their viewing angle markers, round markers, and timestamps, are stored in an image database for subsequent analysis.

[0030] Step S002: Perform frequency domain transformation on the original images under different viewpoints and polishing cycles, and obtain the dark field region under different viewpoints and polishing cycles based on the spatial distribution of high-frequency components in the frequency domain.

[0031] It should be noted that the surface of titanium ingots is uneven. During the polishing process, the raised areas come into contact with the polishing equipment first, while the recessed areas are difficult to reach, resulting in uneven removal of the oxide layer. To address the problem of uneven oxide layer removal on the surface of titanium ingots during polishing, this embodiment analyzes the changes in the texture characteristics of the polished areas and compares them with the texture of the remaining areas, thereby analyzing the proportion of residual oxide layer in each area.

[0032] When the abrasive belt contacts the surface of the titanium ingot, the peeling action of the abrasive grains gradually removes the material layer in the contact area. The oxide layer on the raised areas is ground away first, exposing the underlying titanium alloy body. Since the titanium alloy body and the residual oxide layer have different reflectivities, the former presents a bright field, while the latter presents a dark field, thus creating a clear contrast between bright and dark areas in the image. Based on this, by statistically analyzing the attenuation change of the proportion of low-reflectivity areas (dark field) in different grinding cycles at the same angle, the polishing removal rate at the current viewing angle can be obtained.

[0033] Step S002 further includes steps S0021-S0024: Step S0021: Extract the region of interest from the original images of the titanium ingot under different viewpoints and polishing cycles to obtain the region image.

[0034] Specifically, based on the prior positional relationship between the area where the titanium ingot is located from the current perspective and the industrial camera, the region of interest of the titanium ingot is extracted from the original image.

[0035] For example, in this embodiment, an industrial camera is fixedly mounted directly above the titanium ingot, with the camera's optical axis perpendicular to the center of the ingot's top surface. During the system calibration phase, by capturing an image of the titanium ingot in a standard position (e.g., at a 0° angle on the clamp), the pixel coordinates of the center of the ingot's top surface in the image are determined. , and the pixel length corresponding to the top surface radius. The camera calibration parameters remained constant throughout the polishing process.

[0036] For each acquired raw image (whether before or after polishing), the system extracts the region of interest according to the following steps: read the pre-calibrated center coordinates ( , ) and radius Generate a binary mask with the same size as the original image, where the mask satisfies... The pixels with the specified value are set to 1, and the rest are set to 0. , The coordinates of the pixels are shown below. The original image is multiplied (or bitwise ANDed) with the mask pixel by pixel. In the resulting image, only the pixels within the circular area on the top surface of the titanium ingot are retained, while the pixels outside the area are set to black (grayscale value 0). The final circular area image is the region image.

[0037] Step S0022: Use Fast Fourier Transform to convert the region image into a spectrum image.

[0038] It should be noted that the Fast Fourier Transform (FFT) is a mathematical method for converting an image from the spatial domain to the frequency domain. It decomposes grayscale variations in an image into sinusoidal components of different frequencies. Low-frequency components correspond to regions with gradual grayscale changes (such as large oxide layers or the titanium body itself), while high-frequency components correspond to regions with dramatic grayscale changes (such as polishing scratches, edges, and textures). In this embodiment, the following transformation process is performed on the extracted region image: The region image is converted from a spatial domain grayscale matrix to a complex frequency domain matrix. Specifically, a one-dimensional fast Fourier transform is performed on each row of the image matrix to obtain intermediate results; then a one-dimensional fast Fourier transform is performed on each column of the intermediate results to finally obtain a two-dimensional frequency domain complex matrix.

[0039] The magnitude (i.e., the square root of the sum of the squares of the real and imaginary parts) of each element of the frequency domain complex matrix is ​​calculated to obtain the amplitude spectrum matrix. The numerical values ​​in the amplitude spectrum matrix represent the energy intensity of the corresponding frequency component in the original image.

[0040] The amplitude spectrum matrix is ​​centered (using the FFT shift operation), moving the zero-frequency component to the center of the matrix. This concentrates low-frequency components in the central region of the spectrum, while high-frequency components are distributed around the perimeter. The centered amplitude spectrum is the final output spectrum image.

[0041] Step S0023: Arrange the frequency components in the spectrum image in descending order of their amplitude values ​​to obtain the frequency-amplitude sequence.

[0042] It should be noted that, for subsequent adaptive extraction of the high-frequency portion, the frequency components in the spectrum image need to be sorted by energy level. Specifically, every pixel in the spectrum image except for the DC component (center point) is traversed, and the distance from that point to the center of the spectrum is calculated as its frequency value (unit: period / pixel), and the amplitude value of that pixel (reflecting the energy intensity of that frequency component) is recorded. After ignoring components with amplitude values ​​below a preset noise threshold, all (frequency value, amplitude value) data pairs are arranged in descending order of amplitude value to obtain a frequency. Amplitude sequence. In this sequence, components with larger amplitudes are listed first, corresponding to low-frequency textures with higher energy in the image (such as large areas of oxide layers or titanium body regions); components with smaller amplitudes are listed later, corresponding to high-frequency polishing scratch textures with lower energy.

[0043] Step S0024: Obtain the dark field region under different viewing angles and different polishing cycles based on the frequency values ​​in the frequency-amplitude sequence.

[0044] Step S0024 further includes steps S00241-S00244: Step S00241: Calculate the difference between the frequency values ​​of two adjacent frequency components in the frequency-amplitude sequence, and find the pair of frequency components with the largest absolute value of the difference.

[0045] Specifically, in this embodiment, all adjacent frequency components in the sequence are traversed, the difference between their frequency values ​​is calculated (absolute value is taken), and the pair with the largest difference is found.

[0046] Step S00242: Take the larger frequency value in a pair of frequency components and the largest frequency value in the frequency-amplitude sequence as the interval endpoint, and extract all frequency components within the interval corresponding to the interval endpoint as the high-frequency part.

[0047] Specifically, after finding the pair of frequency components with the largest difference, the component with the larger frequency value in this pair is selected and marked as one endpoint of the high-frequency portion. Then, the entire frequency-amplitude sequence is traversed to find the component with the largest frequency value, which is marked as the other endpoint of the high-frequency portion. These two endpoints are used as the two endpoints of a truncation interval, and all frequency components within that interval are truncated. These truncated frequency components constitute the high-frequency portion of the current spectrum image.

[0048] Step S00243: Perform a fast inverse Fourier transform on the high-frequency component to obtain the reconstructed image.

[0049] It should be noted that the inverse fast Fourier transform is the inverse operation of the fast Fourier transform, used to restore frequency domain data to a spatial domain image.

[0050] The specific process is as follows: First, construct a complex matrix with the same size as the original spectrum image and set all initial elements to zero; then, fill the original coordinate positions in the matrix with the frequency components (including their amplitude and phase information) corresponding to the high-frequency part obtained in the previous step, and keep the remaining low-frequency and unselected frequency components at zero; next, perform an inverse fast Fourier transform on the complex matrix to obtain the complex result; finally, take the real part (or magnitude) of the complex result as the gray value of the reconstructed image.

[0051] Step S00244: The regions with non-zero pixel values ​​in the reconstructed image are identified as polished regions, and the remaining pixel regions are identified as dark areas under different viewing angles and polishing cycles.

[0052] It should be noted that, since only high-frequency components are retained in the reconstructed image, polishing scratches and edges with drastic grayscale changes in the original image will appear as non-zero pixel values ​​in the reconstructed image, while the pixel values ​​of large flat areas (such as unpolished oxide layers or uniform titanium bodies) will approach zero. Therefore, all pixels with non-zero pixel values ​​in the reconstructed image are identified as "polished areas" in this polishing process, indicating that these locations were touched by the polishing tool and caused changes in surface texture; the remaining pixels with zero pixel values ​​are identified as "dark areas," representing that these locations were not significantly affected by this polishing process and still retain their original oxide layer state.

[0053] Step S003: Based on the dark area area under different viewing angles and different polishing cycles, calculate the polishing removal rate under different viewing angles and different polishing cycles.

[0054] It should be noted that: by processing the high-frequency components using inverse Fast Fourier Transform, the pixel region in the current image is obtained and taken as the actual area affected by the polishing process, while the remaining pixel regions in the image are taken as the dark field regions of the oxide layer unaffected by the polishing tool. Thus, based on the degree of change in the surface state of the titanium ingot caused by the polishing equipment during the polishing process, the area affected by the polishing process is accurately extracted.

[0055] Step S003 further includes steps S0031-S0034: Step S0031: Take the dark area under the current viewpoint and the current polishing round as the current dark area, take the dark area under the current viewpoint and the previous polishing round as the previous dark area, and take the dark area at the beginning of the current polishing process as the initial dark area.

[0056] It should be noted that the number of pixels in all dark areas (areas with a pixel value of zero) in the reconstructed image is used as the area of ​​the dark area.

[0057] For each fixed viewpoint, before the polishing process begins (i.e. before the first polishing), an original image is first acquired, the dark field region is extracted according to the aforementioned frequency domain processing method, and its area is calculated. This area is taken as the initial dark field area of ​​this process under that viewpoint, and is recorded as the initial dark field area.

[0058] Subsequently, after each polishing cycle (i.e., each polishing round), the current image from that perspective is acquired. Similarly, the dark field region is extracted through frequency domain processing, and its area is calculated. This area is used as the current dark field area for the current polishing round. At the same time, the system automatically saves the dark field area extracted from the previous polishing round (i.e., after the previous polishing) from that perspective as the previous dark field area.

[0059] Specifically, Indicates the current perspective Current polishing cycle The area of ​​the dark field region below, i.e., the current dark field area. Indicates the current perspective Previous polishing cycle The area of ​​the dark field below is the area of ​​the previous dark field. Indicates the current perspective Current polishing process The initial area of ​​the dark field region.

[0060] Step S0032: Divide the difference between the previous dark field area and the current dark field area by the initial dark field area to obtain the first ratio.

[0061] Specifically, It is denoted as the first ratio.

[0062] Step S0033: The difference between 1 and the first ratio is taken as the polishing removal degree under the current view and the current polishing cycle.

[0063] Specifically, calculate the polishing removal rate under the current viewpoint and the current polishing round:

[0064] in, Indicates the current perspective Current polishing cycle The degree of polishing removal is as follows.

[0065] This represents the reduction rate of the dark field area between adjacent polishing cycles from the current viewing angle. This reduction rate reflects the distribution of the oxide layer on the titanium ingot in the current observation direction and its variation during the polishing process, thus quantifying the degree of oxide layer removal by the amount of reduction in the dark field area.

[0066] Step S0034: Obtain the polishing removal rate under different perspectives and different polishing cycles.

[0067] Specifically, the entire viewpoint is further traversed to obtain the polishing removal degree under each polishing round, and for each individual viewpoint, the polishing removal degree is sorted in ascending order of polishing rounds to obtain the removal degree sequence corresponding to each viewpoint.

[0068] Step S004: Utilize the polishing removal rate under different viewpoints and polishing cycles, as well as the high-frequency components corresponding to the original images under different viewpoints and polishing cycles, to obtain the process completion rate under different viewpoints and polishing cycles.

[0069] It should be noted that the polishing process involves multiple steps, each using grinding equipment with different roughness, achieving a gradual transition from rough polishing to fine polishing. To meet the quality requirements of the finished product, a process switch is required when each step reaches the processing limit of the current equipment, thus achieving intelligent process transition. This adaptive approach allows titanium ingots with different surface qualities to be treated according to their actual processing state, helping to improve the consistency of the surface quality of the finished ingots.

[0070] For the current process, the surface roughness of the grinding equipment determines the resulting surface polishing texture. After a certain number of passes, the current equipment has reached its maximum achievable processing precision, and the improvement in polishing removal rate with each subsequent grinding is very limited. Simultaneously, because the roughness of the grinding equipment remains constant, the similarity between the new texture and existing textures gradually increases. Therefore, by analyzing the similarity of the current scratch frequency distribution and determining whether the current polishing removal rate has approached the processing capacity limit of the current equipment, it can be determined whether the benefit of continuing grinding has significantly decreased.

[0071] Step S004 further includes steps S0041-S0045: Step S0041: Add 1 to the polishing removal degree under the current view and the current polishing round to obtain the sum value, and subtract the polishing removal degree under the current view and the previous polishing round from the sum value to obtain the polishing removal degree improvement amount.

[0072] Specifically, This is recorded as the improvement in polishing removal efficiency. Among them, Indicates the current perspective Current polishing cycle The degree of polishing removal is as follows; Indicates the current perspective Previous polishing cycle The degree of polishing removal is as follows.

[0073] Step S0042: For each frequency value in the high-frequency part under the current viewpoint and the current polishing cycle, calculate the absolute difference between the frequency value and each frequency value in the high-frequency part under the current viewpoint and the previous polishing cycle, and select the minimum absolute difference value.

[0074] Specifically, the high-frequency portion of the current polishing cycle and the high-frequency portion of the previous polishing cycle are obtained by arranging the spectrum image in descending order of amplitude, extracting the high-frequency portion, and then extracting the frequency values ​​of each frequency component.

[0075] After obtaining the high-frequency components of the current polishing cycle and the previous polishing cycle, it is necessary to compare the similarity of the frequency values ​​in these two high-frequency components to determine whether the textures produced by the two polishing processes are similar. The specific procedure is as follows: For each frequency value in the high-frequency range of the current polishing wheel, calculate the absolute difference between it and each frequency value in the high-frequency range of the previous polishing wheel. Among these absolute differences between the current frequency value and all frequency values ​​from the previous polishing wheel, select the smallest value. This smallest value represents the difference between the current frequency value and the closest frequency value from the previous polishing wheel.

[0076] The reason for identifying this minimum difference is that if a high-frequency component in the current polishing cycle is very close to a high-frequency component in the previous polishing cycle (i.e., the minimum difference is small), it indicates that the scratches produced by the two polishing processes are highly consistent in frequency characteristics. This suggests that the polishing equipment in the current process is struggling to generate new textures with different frequencies, meaning its processing capacity is approaching saturation. Conversely, if the minimum difference is large, it indicates that the frequency component was absent or significantly different in the previous polishing cycle, suggesting that the polishing equipment is still generating new texture variations and that there is still processing potential.

[0077] Step S0043: Calculate the mean frequency value of the high-frequency part under the current viewpoint and the current polishing wheel, and divide the minimum absolute difference by the mean frequency value as the polishing intensity similarity of each frequency value in the high-frequency part under the current viewpoint and the current polishing wheel.

[0078] Specifically, Record as the current perspective Current polishing cycle The polishing intensity similarity of the j-th frequency value in the high-frequency part below. Wherein, Indicates the current perspective Current polishing cycle The j-th frequency value in the high-frequency part below; Indicates the current perspective Previous polishing cycle Any frequency value in the high-frequency portion below; This represents the minimum absolute difference. Indicates the current perspective Current polishing cycle The average frequency value of the high-frequency part below.

[0079] Step S0044: Sum the polishing intensity similarity of all frequency values ​​in the high-frequency part under the current view and the current polishing round, and multiply the sum by the polishing removal improvement amount to obtain the process completion degree under the current view and the current polishing round.

[0080] Specifically, calculate the process completion rate under the current viewpoint and the current polishing round:

[0081] in, Indicates the current perspective Current polishing cycle The completion rate of the next process; Indicates the current perspective Current polishing cycle The high-frequency part below The number of frequency values ​​contained in it.

[0082] If the current equipment's processing precision reaches its physical limit, the improvement in polishing removal rate resulting from a single grinding operation... A value approaching 1 indicates that the current polishing process can no longer further reduce the oxide layer. This means that the polishing benefit of the current process has reached its limit, and the continuous changes in surface condition resulting from continuing the current process will become increasingly less noticeable, thus making the timing of process switching more accurate.

[0083] This further characterizes the proximity of the current high-frequency frequencies: among which The minimum value of the difference between the current frequency and the closest frequency from the previous polishing cycle is calculated by comparing the average value of the current high-frequency range with the minimum value of the current frequency. The ratio reflects the similarity between the current polishing mark and the mark from the previous polishing round. When this ratio is small, it indicates that the processing equipment used in the current polishing process can no longer create polishing marks with significant differences, meaning that the current polishing process has reached its maximum processing accuracy peak. At this point, further increasing the processing time will not further improve the processing accuracy, so it should be determined to switch to the next process to ensure that the polishing process proceeds normally and achieves the expected quality.

[0084] Step S0045: Obtain the process completion rate under different perspectives and different polishing cycles.

[0085] Step S005: Determine the process switching weight of the current polishing round based on the process completion degree under different perspectives and polishing rounds.

[0086] It should be noted that: based on the uniformity of process completion under the current perspective, we further evaluate whether the current polishing process has completed its polishing task, and calculate the process switching weight accordingly.

[0087] Taking the rough polishing process as an example: the surface morphology of titanium ingots is complex. The protruding parts are easy to contact with the abrasive belt and be quickly ground flat. From a local perspective, the process completion rate of these areas increases rapidly. However, the concave parts or areas with abnormal hardness are difficult to be fully adhered to by the polishing tool and applied force. The improvement rate of their process completion rate is significantly different from that of areas with higher surface quality, that is, the uniformity is poor.

[0088] If the system determines the current process is complete and switches to the next process based solely on a high overall average completion rate, the cutting force of the fine polishing abrasive (such as fine-grit abrasive belts) will be less than that of the coarse polishing abrasive, resulting in the next process being unable to effectively remove the deep oxide scale or large scratches remaining in the recesses. These untreated surface dead corners will become defects on the final finished ingot, affecting the overall polishing quality.

[0089] Therefore, it is necessary to judge the overall completion status by the uniformity of the overall distribution of process completion from various perspectives, and further determine whether the remaining surface areas need to be processed by the outlier degree of local completion.

[0090] Step S005 further includes steps S0051-S0055: Step S0051: Obtain the process completion degree of all views under the current polishing wheel, and determine the minimum and average process completion degree.

[0091] Specifically, in the current polishing cycle, the process completion degree for each viewpoint has been calculated. The process completion degrees for all views are then aggregated to form a set of completion degree values. The smallest value in this set is identified as the minimum process completion degree; simultaneously, the arithmetic mean of the completion degrees for all views is calculated as the mean process completion degree.

[0092] Step S0052: Divide the minimum process completion rate by the average process completion rate to obtain the degree of non-compliance of the process completion rate for the current polishing round.

[0093] Specifically, Record as the current polishing cycle The degree to which the completion of the process does not meet the standards; This represents the minimum degree of completion of the process. This represents the average completion rate of the process.

[0094] Step S0053: Squaring the difference between the process completion degree and the mean process completion degree of each viewpoint under the current polishing cycle and summing the results, we obtain the first sum value.

[0095] Specifically, This is denoted as the first sum; N represents the number of viewpoints; Step S0054: Divide the number of viewpoints by the first sum to obtain the quotient value, and then take the square root of the quotient value to obtain the overall deviation.

[0096] Specifically, This is recorded as the overall deviation.

[0097] Step S0055: Add 1 to the overall deviation to obtain the second sum value, multiply the degree of non-compliance of the current polishing cycle by the second sum value to obtain the process switching weight of the current polishing cycle.

[0098] Specifically, calculate the process switching weight for the current polishing round:

[0099] in, Indicates the current polishing cycle Process switching weights.

[0100] therefore, This represents the ratio of the worst-case scenario's completion level to the average level in the current round. The closer this ratio is to 1, the closer even the most challenging scenario has been to the overall average polishing level. Furthermore, this is achieved through the overall deviation term... To assess the dispersion of completion scores across different perspectives: when the completion scores of different perspectives are very close, The value is very small. A larger value indicates a larger overall deviation term, suggesting better consistency in polishing quality across different perspectives in the current round. Conversely, a smaller value indicates a smaller deviation term, suggesting insufficient overall uniformity, if the completion levels of different perspectives vary significantly. By multiplying the worst-performing perspective ratio by the overall deviation term and applying a value range limit, the overall compliance of the current process for the surface treatment of titanium alloy ingots can be measured more accurately, providing a reliable basis for process switching.

[0101] Step S006: Based on the process switching weight of the current polishing round, determine whether a polishing process switch is required.

[0102] Step S006 specifically includes: The process switching weight of the current polishing cycle is normalized. If the normalized value is greater than the preset switching threshold, the current polishing process ends and the process switches to the next polishing process; otherwise, the next polishing cycle of the current process continues.

[0103] It should be noted that: due to the weight of process switching The calculation results may exhibit significant numerical fluctuations, making direct comparison with a fixed threshold difficult. Therefore, normalization is necessary. This embodiment uses the Sigmoid function to map the process switching weights to... The interval, the Sigmoid function is in the form of , where parameters The parameter controls the steepness of the mapping curve. This is the center offset, and its specific value can be calibrated through a small number of experiments (e.g., taking...). (Or adjusted based on historical data). The normalized value is the process switching control value for the current polishing cycle.

[0104] The preset switching threshold is This can be obtained through process experimentation or empirical setting based on historical polishing data. When the normalized value is greater than... If the current polishing process has reached its limit, the current polishing process ends and the process switches to the next polishing process (e.g., from rough polishing to medium polishing, or from medium polishing to fine polishing); otherwise, the current process continues to the next polishing cycle. Through this adaptive switching mechanism, each titanium ingot receives a unique polishing cycle allocation based on its own surface condition, thereby ensuring polishing quality while avoiding over-processing.

[0105] After completing the adaptive switching control of the polishing process, in order to further locate the polishing difficulty of each area on the titanium ingot surface, identify residual defects, and guide subsequent rework, this embodiment introduces a construction difficulty index. By quantitatively analyzing the historical performance and relative differences of the process completion from various perspectives, a basis is provided for generating a surface quality heat map and a comprehensive evaluation report.

[0106] The difficulty of polishing the titanium ingot area corresponding to the current viewpoint is assessed by analyzing the distribution of process completion in each polishing process from a single perspective, and whether the process completion of that viewpoint is significantly lower than that of other views.

[0107] The method also includes steps S0071-S0074: Step S0071: Count the number of times the process completion degree is lowest in the current polishing process from the current perspective, and divide the number of times by the number of polishing rounds in the current polishing process to obtain the second ratio.

[0108] It should be noted that the polishing process is a complete processing stage using the same type of grinding tools, while each polishing cycle is an independent grinding action within this process.

[0109] In each polishing cycle, after calculating the process completion degree for all views, the system identifies the view with the lowest completion degree. This view with the lowest completion degree in that cycle is marked as the "worst view" for that cycle. For the current polishing cycle, the system maintains an array of counters, each corresponding to a view. Whenever a view is marked as the worst view in the current cycle, its corresponding counter is incremented by one. After the current polishing cycle ends, the counter value for each view represents the total number of times that view has been identified as having the lowest process completion degree throughout the entire process.

[0110] Specifically, This is denoted as the second ratio; Indicates the current perspective The number of passes with the lowest completion rate in the current polishing process; This indicates the number of polishing passes in the current polishing process.

[0111] Step S0072: Obtain the maximum process completion value of all views under the current polishing cycle, and the standard deviation of the process completion value of the current view in the current polishing process.

[0112] It should be noted that: for the current polishing round, the process completion degree of all perspectives is traversed, and the maximum value is found. At the same time, for the current perspective, the historical process completion degree data of all rounds completed in this polishing process (from the first polishing of this process to the current round) are collected, and the standard deviation of these data is calculated as the standard deviation of the process completion degree of this perspective in the current polishing process.

[0113] Step S0073: Subtract the process completion value of the current view and the current polishing cycle from the maximum process completion value of all views under the current polishing cycle to obtain the difference. Divide the difference by the standard deviation of the process completion value of the current view in the current polishing process to obtain the third ratio.

[0114] Specifically, This is denoted as the third ratio. Indicates the current polishing cycle The maximum completion rate of the process from all perspectives; This represents the standard deviation of the process completion rate from the current perspective in the current polishing process.

[0115] Step S0074: The sum of the second ratio and the third ratio is determined as the construction difficulty from the current perspective.

[0116] Specifically, calculate the construction difficulty from the current perspective:

[0117] in, Indicates the current perspective The construction difficulty.

[0118] The larger the value, the more often this angle becomes a bottleneck in the polishing process due to poor processing results, reflecting the higher processing difficulty of the corresponding area. Based on this, further consideration should be given to the current highest process completion level. Completion rate of the current process from this perspective The difference is compared with the standard deviation of the current viewpoint completion. ratio These two indicators serve as metrics to measure the deviation of the processing from a given perspective. Together, they are used to accurately determine the processing difficulty from each perspective, providing a basis for subsequent quality assessment.

[0119] The method also includes steps S0081-S0083: Step S0081: If the construction difficulty of the current viewpoint is greater than the first safety threshold, then the current viewpoint is determined to be a difficult viewpoint.

[0120] Step S0082: If the construction difficulty of the current view is less than or equal to the first safety threshold and greater than the second safety threshold, then the current view is determined to be the view to be processed.

[0121] Step S0083: If the construction difficulty of the current view is less than or equal to the second safety threshold, then the current view is determined to be a qualified view.

[0122] It should be noted that this embodiment presets two safety thresholds: the first safety threshold... Second security threshold ,and The specific values ​​for both can be determined through process experiments or historical data statistics, for example, by taking... , Among them, the second safety threshold Used to distinguish between "compliance perspective" and "processing perspective": when the construction difficulty is low (≤ When the construction difficulty exceeds a certain threshold, the surface area corresponding to that viewpoint is considered to have met high-quality requirements; when the construction difficulty exceeds a certain threshold, the surface area corresponding to that viewpoint is considered to have met high-quality requirements. But it has not yet exceeded At that time, it was determined that the area still needed further processing. First safety threshold. This is used to identify the "difficulty perspective": once the construction difficulty exceeds... This indicates the presence of severe processing dead zones or stubborn oxide layers in the area, requiring separate marking and rework. Using these two thresholds, the system can classify the surface areas of titanium ingots into three categories, providing a clear basis for thermal mapping and subsequent processing.

[0123] After determining the viewing angle category, the system performs color mapping on the 3D model of the titanium alloy ingot according to the construction difficulty level of each viewing angle: qualified viewing angles (construction difficulty less than or equal to the second safety threshold) are rendered in dark green, viewing angles to be processed (construction difficulty between the second and first safety thresholds) are rendered in yellow, and difficult viewing angles (construction difficulty greater than the first safety threshold) are rendered in dark red. A surface quality heatmap is generated in real time and interactive operation is supported. For high-alarm areas corresponding to difficult viewing angles, the system automatically highlights them in the 3D model and extracts the morphological features and spatial coordinates of the area, ultimately generating a comprehensive quality assessment report. This report includes statistics on overall polishing uniformity, a list of local defects, and specific rework suggestions, thus achieving a complete process from quality assessment to closed-loop feedback.

[0124] In summary, in this embodiment of the invention, by acquiring original images from different perspectives and polishing cycles and extracting dark field regions based on frequency domain transformation, accurate quantitative analysis of the residual oxide layer state on the surface of titanium alloy ingots is achieved. The polishing removal rate is calculated based on the dark field area, and the process completion rate is obtained by integrating the high-frequency component differences between adjacent cycles, thereby enabling the perception of the removal efficiency and texture convergence of the current process. Furthermore, based on the ratio of the minimum to the mean of the process completion rate under all perspectives and the dispersion of the completion rate at each perspective, a switching weight is adaptively determined and compared with a threshold after normalization to trigger process switching. These technical means enable the system to continuously track the polishing process of the same physical area, accurately separate the polishing influence zone and residual dark field at the critical oxide layer removal stage, and automatically determine the switching timing for rough polishing, medium polishing, and fine polishing based on actual processing capabilities. This effectively overcomes the shortcomings of existing technologies, such as difficulties in cross-cycle registration, inaccurate boundary judgment, and reliance on a fixed number of process switching steps, significantly improving the surface quality consistency of titanium alloy ingots while reducing excessive consumption of expensive titanium materials. In addition, heat maps and quality reports are generated based on the construction difficulty, providing intuitive support for defect tracing and closed-loop control.

[0125] This invention also proposes an evaluation system for the surface polishing quality of titanium alloy finished ingots; please refer to [link to relevant documentation]. Figure 2 The diagram illustrates a block diagram of a system for evaluating the surface polishing quality of a titanium alloy ingot, provided in one embodiment of the present invention. The system includes: The acquisition module 100 is used to acquire original images of titanium alloy finished ingots from different angles and under different polishing cycles. Analysis module 200 is used to perform frequency domain transformation on the original images under different viewpoints and polishing cycles, and obtain the dark field region under different viewpoints and polishing cycles based on the spatial distribution of high frequency components in the frequency domain. Based on the area of ​​the dark field region under different viewing angles and different polishing rounds, the polishing removal rate under different viewing angles and different polishing rounds is calculated. By utilizing the polishing removal rate under different perspectives and polishing cycles, as well as the high-frequency components corresponding to the original images under different perspectives and polishing cycles, the process completion rate under different perspectives and polishing cycles can be obtained. The process switching weight for the current polishing round is determined based on the process completion rate under different perspectives and polishing rounds. The determination module 300 is used to determine whether a polishing process needs to be switched based on the process switching weight of the current polishing round.

[0126] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the evaluation system for the surface polishing quality of titanium alloy ingots and the evaluation method for the surface polishing quality of titanium alloy ingots provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0127] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0128] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0129] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for evaluating the surface polishing quality of a finished titanium alloy ingot, characterized in that, The method includes the following steps: Acquire original images of titanium ingots from different perspectives and under different polishing cycles; Frequency domain transformation is performed on the original images from different viewpoints and polishing cycles. Based on the spatial distribution of high-frequency components in the frequency domain, the dark field regions under different viewpoints and polishing cycles are obtained. Based on the dark field area under different viewing angles and different polishing rounds, the polishing removal rate under different viewing angles and different polishing rounds was calculated. By utilizing the polishing removal rate under different viewpoints and polishing rounds, and the high-frequency components corresponding to the original images under different viewpoints and polishing rounds, the process completion rate under different viewpoints and polishing rounds is obtained, specifically including: Add 1 to the polishing removal degree under the current view and the current polishing round to get the sum value, and subtract the polishing removal degree under the current view and the previous polishing round from the sum value to get the polishing removal degree improvement amount; For each frequency value in the high-frequency part under the current viewpoint and the current polishing round, calculate the absolute difference between this frequency value and each frequency value in the high-frequency part under the current viewpoint and the previous polishing round, and select the minimum absolute difference value. Calculate the mean frequency value of the high-frequency part under the current view and the current polishing round, and divide the minimum absolute difference by the mean frequency value to obtain the polishing intensity similarity of each frequency value in the high-frequency part under the current view and the current polishing round; The polishing intensity similarity of all frequency values ​​in the high-frequency part under the current view and the current polishing round is summed, and the sum is multiplied by the polishing removal improvement to obtain the process completion degree under the current view and the current polishing round. Obtain the process completion rate from different perspectives and with different polishing cycles; The process switching weight for the current polishing round is determined based on the process completion rate under different perspectives and polishing rounds, specifically including: Obtain the process completion degree of all views under the current polishing wheel, and determine the minimum and average process completion degree. Divide the minimum process completion rate by the average process completion rate to obtain the degree of non-compliance of the process completion rate for the current polishing round; The first sum is obtained by squaring the difference between the process completion degree and the mean process completion degree of each viewpoint under the current polishing cycle; Divide the number of viewpoints by the first sum to get the quotient, and then take the square root of the quotient to get the overall deviation. Add 1 to the overall deviation to get the second sum. Multiply the degree of non-compliance of the process completion of the current polishing round by the second sum to get the process switching weight of the current polishing round. Based on the process switching weight of the current polishing cycle, determine whether a polishing process switch is required.

2. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 1, characterized in that, The specific steps involved in performing frequency domain transformation on the original images from different viewpoints and polishing cycles, and obtaining the dark field regions from different viewpoints and polishing cycles based on the spatial distribution of high-frequency components in the frequency domain, are as follows: Regions of interest were extracted from the original images of titanium ingots under different viewpoints and polishing cycles to obtain region images; The region image is converted into a spectrum image using Fast Fourier Transform; Arrange the frequency components in the spectrum image in descending order of their amplitude values ​​to obtain the frequency-amplitude sequence. Dark field regions under different viewing angles and polishing cycles are obtained based on frequency values ​​in the frequency-amplitude sequence.

3. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 2, characterized in that, The specific steps for obtaining the dark field region under different viewing angles and polishing cycles based on the frequency values ​​in the frequency-amplitude sequence are as follows: Calculate the difference between the frequency values ​​of two adjacent frequency components in the frequency-amplitude sequence, and find the pair of frequency components with the largest absolute value of the difference; Take the larger frequency value in a pair of frequency components and the largest frequency value in the frequency-amplitude sequence as the interval endpoints, and extract all frequency components within the interval corresponding to the interval endpoints as the high-frequency part. The high-frequency component is subjected to inverse fast Fourier transform to obtain the reconstructed image; The regions with non-zero pixel values ​​in the reconstructed image are identified as polished regions, while the remaining pixel regions are identified as dark areas under different viewing angles and polishing cycles.

4. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 3, characterized in that, The specific steps for calculating the polishing removal rate based on the dark field area under different viewing angles and polishing rounds are as follows: The area of ​​the dark field region under the current view and the current polishing round is taken as the current dark field area, and the area of ​​the dark field region under the current view and the previous polishing round is taken as the previous dark field area. The area of ​​the dark field region at the beginning of the current polishing process is taken as the initial dark field area. The first ratio is obtained by dividing the difference between the previous dark field area and the current dark field area by the initial dark field area. The difference between 1 and the first ratio is taken as the polishing removal degree under the current view and the current polishing round; Obtain the polishing removal rate under different perspectives and different polishing cycles.

5. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 1, characterized in that, The specific steps for determining whether a polishing process switch is needed based on the process switching weight of the current polishing round are as follows: The process switching weight of the current polishing cycle is normalized. If the normalized value is greater than the preset switching threshold, the current polishing process ends and the process switches to the next polishing process; otherwise, the next polishing cycle of the current process continues.

6. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 1, characterized in that, The method further includes: The second ratio is obtained by counting the number of times the process completion rate is lowest from the current perspective in the current polishing process and dividing this number by the number of polishing rounds in the current polishing process. Obtain the maximum process completion rate of all views under the current polishing cycle, and the standard deviation of the process completion rate of the current view in the current polishing process; The difference is obtained by subtracting the process completion value of the current view and the current polishing cycle from the maximum process completion value of all views under the current polishing cycle. The difference is then divided by the standard deviation of the process completion value of the current view in the current polishing process to obtain the third ratio. The sum of the second and third ratios is used to determine the construction difficulty from the current perspective.

7. The method for evaluating the surface polishing quality of titanium alloy finished ingots according to claim 6, characterized in that, The method further includes: If the construction difficulty from the current perspective exceeds the first safety threshold, then the current perspective is determined to be a difficult perspective. If the construction difficulty of the current view is less than or equal to the first safety threshold and greater than the second safety threshold, then the current view is determined to be a view to be processed. If the construction difficulty of the current perspective is less than or equal to the second safety threshold, then the current perspective is determined to be a qualified perspective.

8. A system for evaluating the surface polishing quality of titanium alloy ingots, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for evaluating the surface polishing quality of a titanium alloy finished ingot as described in any one of claims 1-7.

Images