A method for evaluating the sealing degree of a speaker enclosure

Abstract

The application provides a kind of sound box shell sealing degree evaluation method, comprising: reading the corresponding position of compressed mutation gradient and step amplitude combination feature according to the defect coordinate list after refining one by one, grouping similar defect type, determine the leakage severity index of each defect type;With leakage severity index according to numerical interval contrast air-tight grading standard minor, intermediate, serious three threshold value one by one filing, the defect of interval above intermediate is marked as high-risk leakage channel, generate sealing surface air-tight grading result;Sealing surface air-tight grading result and adhesive strip bonding surface three-dimensional scanning image are superimposed, each defect coordinate is mapped on adhesive strip bonding surface corresponding position according to grading result, and sealing defect orientation positioning result is obtained.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for evaluating the sealing degree of a speaker enclosure. Background Technology

[0002] In the modern audio equipment manufacturing industry, the sealing performance of the enclosure directly affects the product's sound quality and lifespan, making it a crucial technical aspect. A good seal not only prevents dust and moisture intrusion but also ensures the stability of the internal acoustic environment, thereby guaranteeing sound purity and equipment durability. Technological advancements in this area have undeniable value in enhancing user experience and product competitiveness. However, a common problem in evaluating and optimizing speaker enclosure sealing performance is an overemphasis on surface phenomena while neglecting the complex influencing factors hidden in the details. Many methods judge sealing effectiveness solely by measuring the gap size at the enclosure seams, lacking effective identification and countermeasures for the interference of unavoidable minor imperfections during enclosure manufacturing on overall airtightness. The tiny flash that may exist at the enclosure seams—microscopic protrusions left over from the injection molding process—has a profound impact on airtightness due to its angle of intersection with the sealing material. If the flash penetrates the sealing material at a perpendicular or near-perpendicular angle, it creates a small overhead channel, allowing air or moisture to leak through. Even with a very low flash height, this phenomenon can lead to seal failure because the presence of channels disrupts the uniform adhesion of the sealing material, creating a directional leakage path. Specifically, in the production of speaker housings, the injection molding process often leaves tiny flashes at the parting line of the housing. This is because the parting line itself is the closed joint of the upper and lower molds. During the injection molding and holding pressure stage, the molten plastic is pushed by high pressure into the tiny gap between the closed surfaces of the molds. After cooling and solidification, a ridge-like plastic protrusion remains along the parting line. The sealing strip relies on the elastic deformation of its own rubber material to continuously press against the housing surface along its direction to achieve a seal. When the direction of the aforementioned ridge-like flash happens to intersect the direction of the sealing strip at a large angle, the flash protrusion will push up a part of the sealing strip at a certain cross-section along the length of the strip. Due to its own compression and rebound capacity, the sealing strip cannot completely backfill and adhere along the steeply descending sidewalls on both sides of the flash, leaving a small gap running through the width of the sealing strip on the surface of the strip along the flash direction. Taking a real-world example, at the junction of the upper and lower casings of a speaker, if the flash cuts into the area where the sealing strip is installed at a near-vertical angle, even if the flash height is only a few tens of micrometers, it is enough to prevent the sealing strip from adhering completely, forming a hidden leakage channel. This can lead to problems such as moisture absorption of internal components or sound quality distortion in high-humidity environments. Therefore, accurately identifying the tiny gaps caused by the intersection angle between the flash and the sealing material, and quantifying their specific impact on airtightness, has become a key issue in optimizing the sealing performance of speaker casings. Summary of the Invention

[0003] This invention provides a method for evaluating the sealing degree of a speaker enclosure, mainly including: Acquire three-dimensional scanning images of the adhesive strip bonding surface, coordinates of the parting line path, and pressure readings of the diaphragm during the pressure test in the sealed cavity. Extract the flash profile based on the three-dimensional scanning images of the adhesive strip bonding surface and calculate the intersection angle between the flash and the adhesive strip direction. Based on the intersection angle between the flash and the direction of the rubber strip, potential linear channel risk areas are identified and marked. Their spatial range is mapped to the corresponding coordinates of the pressure readings. Local abrupt changes in the pressure readings within this range are analyzed and converted into compression abrupt change gradients. The pressure readings of the diaphragm during the pressurization test in the sealed cavity are fitted along the time axis to generate an overall leakage rate curve of the sealing surface. The time and magnitude of the step in the curve are identified, and the spatial coordinates of the step point and the potential linear channel risk area are associated to obtain the initial location coordinate set of directional leakage. Based on the initial positioning coordinate set and the parting line path coordinates, the spatial position of each candidate defect is recalibrated to obtain its precise spatial coordinates in the area where the adhesive strip and the parting line intersect. The relationship between the actual direction of the flash and the adhesive strip path is compared at this coordinate. The coordinate landing points of the linear channel that forms a vertical passage through the adhesive strip path are screened and added to the refined defect coordinate list. Read the combined characteristics of compression mutation gradient and step amplitude at the corresponding position one by one according to the refined defect coordinate list, group similar defect types, and determine the leakage severity index of each defect type. The severity of leakage indicators are filed one by one according to the three thresholds of slight, moderate and severe in the airtightness classification standard according to the numerical range. Defects in the moderate and above ranges are marked as high-risk leakage channels, and the airtightness classification results of the sealing surface are generated. The airtightness classification results are superimposed with the three-dimensional scan image to obtain the location result of the sealing defect.

[0004] Furthermore, the burr contour is extracted from the 3D scan image of the adhesive strip bonding surface, and the intersection angle between the burr and the adhesive strip direction is calculated, including: Surface point cloud data is extracted from the three-dimensional scanned image, and cross-sectional slices are generated along the boundary of the adhesive strip application area. Extract the surface height value of each slice, and mark the area where the height difference between adjacent slices exceeds the preset benchmark as a convex candidate area; By vectorizing and connecting the edges of the candidate protrusion areas, the spatial orientation curve of the flash profile is obtained; Generate a reference baseline based on the path coordinates of the parting line, and read the laying direction of the adhesive strip; The burr outline is projected onto the bonding surface to form a projection line segment. The angle between the projection line segment and the direction of the adhesive strip is measured to obtain the intersection angle value.

[0005] Furthermore, potential linear channel risk areas are identified and marked based on the intersection angle between the flash and the adhesive strip direction. Their spatial range is mapped to the corresponding coordinates of the pressure readings. Local abrupt changes in pressure readings within this range are analyzed and converted into compression gradient changes, including: Filter out the projected line segments whose intersection angle exceeds a preset vertical threshold, and mark the covered area as a potential linear channel risk zone; Extract the boundary coordinate range of the risk zone and establish a mapping relationship between the coordinates of the adhesive strip bonding surface and the coordinates of the pressure distribution data grid; The risk zone boundary coordinates are converted into grid row and column index intervals to obtain the coordinate mapping result; For the grid interval defined by the coordinate mapping result, read the pressure value of each grid node; The pressure distribution is scanned along the width of the adhesive strip to identify the locations where the pressure difference between adjacent grid nodes exceeds the preset mutation threshold. The locations of local mutations are determined, and the ratio of pressure difference to grid spacing is calculated to obtain the compression mutation gradient.

[0006] Furthermore, the pressure readings of the diaphragm during the pressurization test in the sealed cavity are fitted along the time axis to generate an overall leakage rate curve of the sealing surface. The timing and magnitude of the step in the curve are identified, including: For each grid node, the pressure change trajectory is plotted, and the decay rate is extracted using a fitting method. Summarize the decay rates of each node to generate an overall leakage rate curve for the sealing surface; Scan along the time axis of the leakage rate curve to identify locations where the pressure difference between adjacent sampling points exceeds a preset change threshold, and mark them as step moments; Read the pressure drop before and after the step moment and record the step amplitude.

[0007] Furthermore, the spatial coordinates of the associated step points and the potential linear channel risk area are used to obtain the initial location coordinate set of the directional leakage, including: sorting and filtering the risk-associated step points by step amplitude, mapping the grid coordinates to the sealing surface, and summarizing to obtain the initial location coordinate set of the directional leakage.

[0008] Furthermore, based on the initial positioning coordinate set and the parting line path coordinates, the spatial position of each candidate defect is recalibrated to obtain its precise spatial coordinates in the area where the adhesive strip bonding surface intersects with the parting line. At these coordinates, the relationship between the actual direction of the flash and the adhesive strip path is compared. The coordinate points that constitute a linear channel perpendicularly traversing the adhesive strip path are then filtered and added to the refined defect coordinate list, including: The starting positions of each candidate defect in the initial positioning coordinate set are obtained. The nearest node is found along the path coordinates of the parting line as a calibration reference point. The precise spatial coordinates are obtained by correcting the offset along the normal direction. Extract the vectors of the actual direction of the burr and the path of the adhesive strip, calculate the angle between the actual direction of the burr and the path of the adhesive strip, and determine that a vertical crossing channel is formed when the angle is within the vertical threshold range and mark the effective defect landing point. Based on the distribution location, redundant items that are too close to each other are sorted and redundant items are removed. The effective defect landing points are recorded to obtain a list of defect coordinates.

[0009] Furthermore, the coordinate points of the linear channels that constitute the vertical path of the adhesive strip are filtered and imported into the refined defect coordinate list, including: Extract the calibration coordinates of the area where the adhesive strip mating surface intersects with the parting line, read the flash outline and the sealing laying direction, and obtain the included angle by sampling the direction vector pairing; If the included angle is within the threshold, the coordinates are marked as vertical entry points. When the number of consecutive vertical entry points exceeds the continuity threshold, the coordinates are determined to be candidate defect landing points. Arrange and archive the defects according to their spatial location on the sealing surface, filter out the candidate defect locations that meet the preset conditions, and obtain the refined defect coordinate list.

[0010] Furthermore, based on the refined defect coordinate list, the combined characteristics of the compression abrupt change gradient and step amplitude at the corresponding locations are read one by one. Similar defect types are grouped, and the leakage severity index for each defect type is determined, including: Read the compression abrupt gradient and step amplitude of each point in the list of refining defect coordinates, and pair them to form a combined feature vector; Clustering is used to measure similarity by feature vector distance. Points with a distance less than a preset threshold are grouped into the same group to obtain defect type groups. For each defect type group, the mean values ​​of step amplitude and compression mutation gradient within each group are calculated. The leakage severity index for each defect type is obtained by weighting and summing the mean values ​​of the step amplitude and compression mutation gradient within each group according to preset weights.

[0011] Furthermore, the severity of the leakage is categorized into three threshold levels—minor, moderate, and severe—based on numerical ranges and compared to the airtightness grading standard. Defects in the moderate and above ranges are marked as high-risk leakage channels, generating an airtightness grading result for the sealing surface, including: The severity index of the leak is obtained and compared with the three threshold ranges of the airtightness classification standard. If the leak falls into the corresponding range, it is classified as slight, moderate or severe, and an airtightness level label is obtained. The airtightness rating labels are selected as medium or severe, and all corresponding defect locations are marked as high-risk leakage channels. The airtightness rating labels and marking status are combined to obtain the airtightness classification result of the sealing surface.

[0012] Furthermore, the airtightness grading results of the sealing surface are overlaid with the three-dimensional scan image of the adhesive strip bonding surface. The coordinates of each defect are mapped to the corresponding position on the adhesive strip bonding surface according to the grading results to obtain the location results of the sealing defects, including: A reference system is established by reading the spatial coordinates of the three-dimensional scanned image of the adhesive strip bonding surface, and the coordinates of the defect landing point are transformed to obtain the mapped coordinates. The airtightness level label and high-risk status are superimposed at the corresponding mapping coordinates. Different symbols are used to distinguish the airtightness levels corresponding to the three thresholds. High-risk leakage channels are marked with highlighted labels to obtain the superimposed image. The spatial location of the defect and the airtightness information are arranged in order to obtain the location result of the sealing defect.

[0013] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method for evaluating the sealing degree of a speaker housing. Addressing the problem of sealing failure caused by burr defects at the adhesive strip bonding surface, this invention innovatively integrates three-dimensional scanning, pressure distribution analysis, and leakage path localization techniques. The invention acquires a three-dimensional scan image of the adhesive strip bonding surface, extracts the burr contour, calculates its intersection angle with the adhesive strip direction, and identifies potential linear channel risk areas. Combining pressure distribution acquisition with time-axis attenuation rate fitting, a leakage rate curve is generated, accurately locating the spatial correlation between step points and risk areas. Further, defect coordinates are calibrated, linear channels perpendicularly crossing the adhesive strip path are screened, the defect list is refined, and the severity of leakage is assessed by compressing abrupt gradients and step amplitudes. Finally, an airtightness grading result is generated and mapped onto the bonding surface image, achieving precise defect location. This invention significantly improves the accuracy and efficiency of speaker housing sealing defect detection, providing a reliable basis for product quality control. Attached Figure Description

[0014] Figure 1 This is a flowchart of a method for evaluating the sealing degree of a speaker housing according to the present invention.

[0015] Figure 2 This is a schematic diagram of a method for evaluating the sealing degree of a speaker housing according to the present invention.

[0016] Figure 3 This is another schematic diagram of a method for evaluating the sealing degree of a speaker housing according to the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] like Figures 1-3 This embodiment of a method for evaluating the sealing degree of a speaker housing may specifically include: Step S101: Obtain a three-dimensional scan image of the adhesive strip bonding surface, the coordinates of the parting line path, and the pressure readings of the diaphragm during the pressure test in the sealed cavity. Extract the flash profile based on the three-dimensional scan image of the adhesive strip bonding surface and calculate the intersection angle between the flash and the adhesive strip direction.

[0019] A 3D scan image of the adhesive strip bonding surface is acquired. Surface point cloud data is extracted from the 3D scan image. Cross-sectional slices are generated layer by layer along the boundary of the adhesive strip laying area. The surface height value of each slice is extracted. Areas where the height difference between adjacent slices exceeds a preset reference height are marked as candidate protrusion areas. Vectorized lines are drawn along the edges of the candidate protrusion areas to obtain the spatial direction curve of the flash profile. Based on the parting line path coordinates, a reference baseline for the parting line is generated on the adhesive strip bonding surface. The spatial direction curve of the flash profile is aligned with the reference baseline to calibrate the position. The laying direction of the adhesive strip is read. The spatial direction curve of the flash profile is projected onto the bonding surface plane to form flash projection line segments. The angle between the flash projection line segment and the adhesive strip laying direction is measured segment by segment to obtain the corresponding intersection angle value of each line segment. Pressure readings of the diaphragm during the pressurization test in the sealed cavity were collected based on the pressure distribution. These pressure readings were aligned with the diaphragm grid coordinates at the sampling time. Within the coordinate range of the mating surface corresponding to the projection line segment of the flash where the intersection angle exceeds a preset angle threshold of 45 degrees, local pressure values ​​were extracted and compared with the pressure values ​​of the surrounding area. The surrounding area was defined as a circular region with a radius of 5 cm centered at this coordinate point. The difference was calculated using the formula D = |P| local -avg(P surround )|, where D is the difference magnitude, P local For local pressure, avg(P) surround The average pressure of the surrounding area is used to pair the difference magnitude with the corresponding cross angle value and output the cross angle between the burr and the direction of the rubber strip.

[0020] During the speaker enclosure sealing performance testing, the acquisition of 3D scan images of the adhesive strip bonding surface is achieved using a structured light 3D scanner to collect data across the entire surface of the enclosure joint area. The scanner projects a striped grating onto the surface of the adhesive strip application area, and the depth coordinates of each sampling point are calculated using the phase shift method, forming a 3D point cloud data set covering the entire bonding surface. This point cloud data set contains the spatial 3D coordinates of each sampling point on the bonding surface. The spacing between sampling points is set according to the scanning accuracy, and higher density sampling is used in the area near the speaker enclosure parting line to capture detailed features of the flash.

[0021] Specifically, when generating cross-sectional slices layer by layer along the boundary of the adhesive strip laying area, the width direction of the adhesive strip is taken as the normal direction of the slice, and several cross-sections are sequentially cut in the laying area at preset intervals. Each cross-section intersects with the point cloud data to form a contour curve, and the height value of each point on the contour curve is extracted as the surface height data sequence of the slice.

[0022] In one embodiment, the height difference between adjacent slices is calculated using a point-to-point correspondence method. The height values ​​of the previous slice and the next slice at the same lateral position are subtracted. If the difference exceeds a preset reference height, it is determined that there is a surface protrusion at that position, and the area corresponding to the protrusion is the candidate area for the burr protrusion.

[0023] For example, the preset reference height is determined based on the compression rebound of the adhesive strip and the injection molding process tolerance of the housing. When the height difference exceeds this reference value, it indicates that the protrusion at that location is sufficient to affect the uniform adhesion of the adhesive strip. The candidate protrusion area appears as a strip-shaped region distributed along the parting line on the bonding surface. Its edge position is formed by connecting the lateral coordinates of the protrusion points in adjacent slices. The edge coordinate points are vectorized and connected using spline interpolation to obtain the spatial direction curve of the flash profile. This curve reflects the extension direction and geometry of the flash on the bonding surface.

[0024] It should be noted that the parting line path coordinates are extracted from the shell mold design data, representing the projection trajectory of the closed joint of the upper and lower molds on the shell surface. A parting line reference baseline is generated on the adhesive strip bonding surface based on these path coordinates. This baseline serves as the geometric benchmark for comparing the flash direction with the adhesive strip laying direction.

[0025] In one possible implementation, the spatial orientation curve of the flash profile is orthographically projected onto the bonding surface plane, and the height component is eliminated to form a flash projection line segment. When measuring the angle between the flash projection line segment and the adhesive strip laying direction segment by segment, the tangent direction of the adhesive strip laying direction is used as the reference vector, and the direction of the flash projection line segment is used as the measured vector. The angle value between the two vectors is calculated as the intersection angle value corresponding to the line segment. During the pressure distribution acquisition test of the diaphragm in the sealed cavity, the pressure readings of each grid node are recorded at fixed time intervals. After aligning the pressure readings with the diaphragm grid coordinates according to the sampling time, local pressure values ​​are extracted within the bonding surface coordinate range corresponding to the flash projection line segment where the intersection angle value exceeds a preset angle threshold. By comparing the pressure value of this local area with the pressure value of the surrounding area unaffected by the flash, local abnormal features of the pressure distribution are identified. The difference amplitude is paired and labeled with the intersection angle value of the corresponding line segment to obtain the intersection angle between the flash and the adhesive strip direction and its response correlation at the pressure distribution level.

[0026] Step S102: Identify and mark potential linear channel risk areas based on the intersection angle between the flash and the direction of the adhesive strip, map their spatial range to the corresponding coordinates of the pressure reading, analyze the local abrupt changes in the pressure reading within this range, and convert them into compression abrupt change gradients.

[0027] Based on the intersection angle between the burr and the adhesive strip direction, burr projection segments with intersection angle values ​​exceeding a preset vertical angle threshold of 85 degrees are selected. The adhesive strip bonding surface area covered by these segments is marked as a potential linear channel risk zone. The boundary coordinate range of the risk zone is extracted. Specifically, the endpoint coordinates A(x1, y1) and B(x2, y2) of the segment are first identified. Then, a preset influence distance d=5 mm is extended along the adhesive strip direction to obtain the extended boundary. That is, the new coordinate range is x from min(x1, x2)-d to max(x1, x2)+d, and y from min(y1, y2)-d to max(y1, y2)+d, thus obtaining the spatial range definition of each potential linear channel risk zone. The grid coordinates of the pressure distribution acquisition diaphragm and the spatial range definition of the potential linear channel risk zone are obtained. A mapping relationship between the adhesive strip bonding surface coordinates and the diaphragm grid coordinates is established. The boundary coordinate range of each risk zone is converted into the corresponding row and column index intervals on the diaphragm grid to obtain the coordinate mapping result of the risk zone at the pressure reading acquisition level. For the diaphragm grid interval defined by the coordinate mapping result, the pressure readings of each grid node in the interval are read. The numerical distribution of the pressure readings is scanned row by row along the width of the adhesive strip. The location where the pressure reading difference between adjacent grid nodes exceeds the preset abrupt change threshold of 2 kPa is identified as a local abrupt change location. The ratio of the pressure reading difference at the local abrupt change location to the grid spacing is calculated. Specifically, the grid spacing s is in millimeters. The pressures of adjacent nodes p1 and p2 are input, and the gradient g=(p2-p1) / s is output. For example, when p1=10 kPa, p2=15 kPa, and s=2 mm, g=2.5 kPa / mm, and the compression abrupt change gradient of that location is obtained.

[0028] During the speaker enclosure sealing performance testing, the intersection angle between the burr and the adhesive strip direction is derived from the calculation results of the angle between the burr projection line segment and the adhesive strip laying direction in the previous step. Based on the intersection angle value, burr projection line segments that exceed a preset vertical angle threshold are selected, and the adhesive strip bonding area covered by such line segments is marked as a potential linear channel risk area.

[0029] Specifically, the potential linear channel risk zone refers to the area where the burr cuts into the adhesive strip laying path at a near-vertical angle, and the burr protrusion within this area has the geometric conditions to form an overhead leakage channel. When extracting the boundary coordinate range of each risk zone, a preset influence distance is extended along both ends of the burr projection line segment towards the width of the adhesive strip to form a rectangular boundary covering the area traversed by the burr, thus obtaining the spatial range definition of each potential linear channel risk zone.

[0030] In one embodiment, when establishing the mapping relationship between the coordinates of the adhesive strip bonding surface and the diaphragm grid coordinates, the geometric correspondence between the positioning reference point during diaphragm installation and the adhesive strip bonding surface is obtained. Based on the row and column spacing of the diaphragm grid and the physical dimensions of the adhesive strip bonding surface, the number of grid nodes per unit length of the bonding surface is calculated, and the boundary coordinate range of the risk zone is converted into a row and column index interval on the diaphragm grid. The coordinate mapping result defines the specific location range of each risk zone at the pressure reading acquisition level, enabling subsequent identification of local abrupt changes in pressure readings to focus on specific areas related to flash crossing.

[0031] It should be noted that when scanning pressure readings within the diaphragm grid interval defined by the coordinate mapping results, the pressure values ​​recorded by each grid node within this interval during the pressurization test are read. The distribution of pressure reading values ​​is scanned row by row along the width of the adhesive strip, and the difference in pressure readings between adjacent grid nodes is calculated. Furthermore, locations where the pressure reading difference between adjacent grid nodes exceeds a preset abrupt change threshold are identified as local abrupt change locations. These local abrupt change locations indicate that the adhesive strip is experiencing uneven compression due to flash at that location. The ratio of the pressure reading difference at the local abrupt change location to the grid spacing is calculated to obtain the compression abrupt change gradient at that location. This gradient value reflects the degree of drastic change in the amount of adhesive strip compression along the spatial direction.

[0032] Step S103: Fit the attenuation rate of the pressure readings of the pressure distribution acquisition diaphragm during the pressurization test in the sealed cavity along the time axis to generate an overall leakage rate curve of the sealing surface. Identify the time and magnitude of the step in the curve, associate the step point with the spatial coordinates of the potential linear channel risk area, and obtain the initial location coordinate set of directional leakage.

[0033] The pressure reading sequence recorded by the pressure distribution acquisition diaphragm during the pressurization test in the sealed cavity is obtained. The pressure readings are arranged along the time axis according to the sampling time. For the pressure reading of each diaphragm grid node, the pressure change trajectory is plotted along the time axis. The least squares method is used to linearly fit the pressure change trajectory of each node, and the slope of the fitted line is extracted as the decay rate of that node. The decay rates of all nodes are summarized according to their spatial position on the diaphragm to obtain the overall leakage rate curve of the sealing surface. The overall leakage rate curve of the sealing surface is scanned point by point along the time axis to identify the positions on the curve where the pressure value difference between adjacent sampling points exceeds a preset change threshold. The sampling time corresponding to the position is marked as the step time. The absolute value of the pressure value before and after the step time is read as the step amplitude. Each step time and its corresponding step amplitude are paired and recorded to obtain a set of step feature points. Based on the spatial coordinate range of the potential linear channel risk zone, the diaphragm grid node coordinates corresponding to each step moment in the set of step feature points are compared with the spatial coordinates of the risk zone. If the grid node coordinates corresponding to the step moment fall within the spatial coordinate range of the risk zone, the step feature point is marked as a risk-related step point, and the grid coordinates of the risk-related step point are obtained. The grid coordinates of the risk-related step points are sorted from largest to smallest step amplitude, and risk-related step points with step amplitudes exceeding a preset amplitude threshold are filtered. Based on the positional mapping relationship between the diaphragm grid and the sealing surface, the filtered grid coordinates are converted into sealing surface coordinates. The sealing surface coordinates of each step point and their corresponding step amplitudes are collected to obtain the initial location coordinate set for directional leakage.

[0034] During the speaker enclosure sealing performance testing, a pressure distribution acquisition diaphragm covers the adhesive strip bonding area of ​​the sealing cavity. The diaphragm surface is divided into several grid nodes, each equipped with a pressure sensing unit. During the pressure test of the sealing cavity, each grid node continuously records the pressure value according to a preset sampling frequency, forming a pressure reading sequence indexed by the sampling time.

[0035] Specifically, the pressure readings for each diaphragm grid node are arranged with the sampling time on the horizontal axis and the pressure value on the vertical axis to form a pressure change trajectory for that node. This trajectory reflects the evolution of pressure at that node over time during the pressurization test. In well-sealed areas, the pressure change trajectory shows a stable characteristic, while in areas with leakage channels, the pressure change trajectory shows a continuous downward trend.

[0036] In one embodiment, the least squares method is used to linearly fit the pressure change trajectory of each node. The least squares method finds a straight line that minimizes the sum of the squares of the perpendicular distances from all sampling points to that line, thus obtaining the slope and intercept of the fitted line. In a sealing performance testing scenario, the slope of the fitted line characterizes the rate at which the pressure decreases over time; the absolute value of this slope is the decay rate of that node. A larger decay rate indicates faster pressure dissipation at that node location, and a higher probability of a leakage path. The decay rates of each grid node are summarized and arranged according to their spatial position on the diaphragm to form a leakage rate distribution covering the entire sealing surface. The overall leakage rate curve of the sealing surface is then plotted with the time axis as the reference.

[0037] It should be noted that the overall leakage rate curve of the sealing surface is plotted with the sampling time as the horizontal axis and the spatial mean or weighted mean of the decay rate of each node as the vertical axis. When scanning point by point along the time axis of this curve, the position where the pressure difference between adjacent sampling points exceeds the preset change threshold is identified, and this position is the moment when the step occurs.

[0038] For example, the determination of a step moment is based on whether the difference between the pressure values ​​of two adjacent sampling points exceeds a preset change threshold. If the absolute value of the difference between the pressure values ​​of the previous and subsequent sampling points exceeds this threshold, the time corresponding to the subsequent sampling point is marked as a step moment. The step amplitude is defined as the absolute value of the difference in pressure values ​​before and after the step moment, reflecting the severity of the pressure change at that moment. Each step moment is paired and recorded with its corresponding step amplitude to form a set of step feature points. Each step feature point in this set contains a time marker and amplitude value of the step occurrence. Further, based on the spatial coordinate range of the potential linear channel risk area obtained previously, the coordinates of the membrane grid nodes corresponding to each step moment in the set of step feature points are compared with the spatial coordinates of the risk area. If the grid node coordinates corresponding to the step moment fall within the spatial coordinate range of the risk area, it indicates that the step feature point is spatially associated with the potential linear channel formed by the burr crossing the adhesive strip, and the step feature point is marked as a risk-associated step point.

[0039] In one possible implementation, the grid coordinates of the risk-related step points are sorted from largest to smallest step amplitude, and risk-related step points with step amplitudes exceeding a preset amplitude threshold are filtered out. The preset amplitude threshold is determined based on sealing performance standards and historical test data, and is used to filter out minor fluctuations with small amplitudes. Based on the positional mapping relationship between the diaphragm grid and the sealing surface, the grid coordinates of the filtered risk-related step points are converted into actual coordinates on the sealing surface. This mapping relationship is established based on the positioning reference during diaphragm installation and the geometric dimensions of the sealing surface, achieving a correspondence between the grid row and column indices and the physical location of the sealing surface. The sealing surface coordinates of each step point and its corresponding step amplitude are collected to obtain an initial location coordinate set for directional leakage. This coordinate set marks the specific location on the sealing surface where there is a risk of directional leakage and the severity of the leakage.

[0040] Another way to determine the initial location coordinate set of directional leakage is to retrieve the time markers and amplitude values ​​corresponding to each step point, compare them one by one with the boundary range of the potential linear channel risk area, identify the leakage trigger point that is synchronized with the pressure response of the risk area at the time of the step occurrence, determine the actual position of the trigger point on the sealing surface and calibrate it to form the initial location coordinate set of directional leakage.

[0041] The time markers and amplitude values ​​corresponding to each step point are retrieved, and the boundary coordinates of the potential linear channel risk zone are read. The membrane grid coordinates of each step point are compared with the risk zone boundary range one by one, and step points whose membrane grid coordinates fall within the risk zone boundary range are selected to obtain a set of candidate step points located within the risk zone. For each step point in the candidate step point set, its step occurrence time and corresponding amplitude value are extracted. The pressure change amplitude of the spatially adjacent grid nodes on the membrane grid at the same time is read synchronously. If the spatially adjacent grid nodes also experience a sudden pressure drop at the same time, and the difference between the drop amplitude and the amplitude value of the step point is within a preset synchronization threshold range, then the step point is determined to be a leakage trigger point, and the time marker and grid coordinates of the leakage trigger point are obtained. Based on the grid coordinates of the leakage trigger point and the position mapping relationship between the diaphragm and the sealing surface, the grid coordinates of the leakage trigger point are converted into the actual position coordinates on the sealing surface. The leakage direction is determined according to the arrangement direction between the leakage trigger point and the grid nodes that are spatially adjacent and synchronously responding. The actual position coordinates of the leakage trigger point are marked as the starting position. The leakage direction and starting position of each leakage trigger point are collected to form a set of initial positioning coordinates for directional leakage that includes the leakage direction and the starting position.

[0042] During the speaker enclosure sealing performance testing, the time markers and amplitude values ​​corresponding to each step point are derived from the scanning and identification results of the overall leakage rate curve of the sealing surface in the previous step. When retrieving these step point data, the boundary range coordinates of the potential linear channel risk area are read simultaneously. These boundary range coordinates are determined by the area where the angle between the flash and the direction of the rubber strip exceeds a preset angle threshold.

[0043] Specifically, when comparing the membrane grid coordinates of each step point with the boundary range of the risk zone one by one, it is determined whether the grid row and column index of the step point falls within the row and column index interval defined by the risk zone boundary. If the grid coordinates of the step point are within this interval, the step point is included in the candidate step point set, indicating that the pressure change location of the step point spatially overlaps with the potential linear channel formed by the flash crossing the rubber strip.

[0044] In one embodiment, pressure response synchronization is determined for each step point in the candidate step point set. The spatially adjacent grid nodes refer to the four grid nodes (up, down, left, and right) that share a boundary with the step point on the diaphragm grid. After extracting the step occurrence time of the step point, the pressure change amplitudes of these four spatially adjacent grid nodes at the same time are read, and the difference between the pressure change amplitudes of each adjacent node and the amplitude value of the step point is calculated. If the absolute value of the difference is within a preset synchronization threshold range, it is determined that the adjacent node and the step point exhibit synchronized characteristics in pressure response, and thus the step point is determined to be a leakage trigger point. The preset synchronization threshold range is determined based on the material elasticity of the sealing strip and the cavity pressurization rate, and is used to identify multi-node coordinated pressure drop phenomena caused by the same leakage channel.

[0045] It should be noted that when converting the grid coordinates of the leakage trigger point to the actual position coordinates of the sealing surface, a mapping relationship is established between the positioning reference point recorded during diaphragm installation and the geometric dimensions of the sealing surface, and the grid row and column indices are converted into physical distance coordinates on the sealing surface. Furthermore, the leakage direction is determined based on the arrangement direction between the leakage trigger point and its spatially adjacent and synchronously responding grid nodes.

[0046] For example, if the adjacent node to the right of the leak trigger point exhibits a synchronous pressure drop while the node to the left does not, the leak direction is determined to be extending from left to right. The actual position coordinates of the leak trigger point are marked as the starting position, and the leak direction and starting position of each leak trigger point are collected to form a set of initial location coordinates for directional leaks that includes the leak direction and starting position.

[0047] Step S104: Based on the initial positioning coordinate set and the parting line path coordinates, the spatial position of each candidate defect is recalibrated to obtain its precise spatial coordinates in the area where the adhesive strip bonding surface and the parting line intersect. The relationship between the actual direction of the flash and the adhesive strip path is compared at these coordinates, and the coordinate landing points that constitute the linear channel that vertically crosses the adhesive strip path are screened and incorporated into the refined defect coordinate list.

[0048] The starting position coordinates of each candidate defect in the initial location coordinate set of the directional leakage are obtained. Simultaneously, the parting line path coordinates are read. For each candidate defect's starting position coordinates, the nearest parting line node along the parting line path coordinates is found as a calibration reference point. The distance offset between the candidate defect's starting position and the calibration reference point is calculated. Based on this offset, the starting position coordinates of the candidate defect are corrected along the normal direction of the parting line to obtain the calibrated candidate defect coordinates. Based on the calibrated candidate defect coordinates, the area where these coordinates are located is located on the adhesive strip bonding surface. The direction vector of the adhesive strip laying path and the direction vector of the parting line within this area are extracted. The boundary range of the intersection area between the adhesive strip bonding surface and the parting line is determined based on these two direction vectors, obtaining the precise spatial coordinates of each candidate defect within the intersection area. For the precise spatial coordinates, the actual curve segment of the burr contour at that coordinate is read, and the direction vector of the curve segment is extracted. Simultaneously, the laying direction vector of the adhesive strip path at that coordinate is read, and the angle between the burr direction vector and the adhesive strip laying direction vector is calculated. If the angle is within a preset vertical angle threshold range, it is determined that the burr at that coordinate constitutes a linear channel perpendicularly crossing the adhesive strip path, and the coordinate is marked as a valid defect landing point. The precise spatial coordinates of the valid defect landing points and their corresponding burr direction information are collected and sorted according to the distribution position of each valid defect landing point on the adhesive strip bonding surface. Redundant landing points with a distance less than a preset distance threshold between adjacent landing points are removed. The filtered valid defect landing points and their coordinate information are then incorporated into a refined defect coordinate list.

[0049] During the speaker enclosure sealing performance testing, the starting position coordinates of each candidate defect in the initial location coordinate set for directional leakage are derived from the previous correlation identification results of step feature points and risk area spatial coordinates. After obtaining these starting position coordinates, the parting line path coordinates are read simultaneously. These path coordinates record the complete trajectory node sequence of the parting line on the adhesive strip bonding surface.

[0050] Specifically, for each candidate defect's starting position coordinates, all nodes along the parting line's path coordinates are traversed, the Euclidean distance between the starting position and each node is calculated, and the node with the smallest distance is selected as the calibration reference point. This calibration reference point represents the position on the parting line closest to the candidate defect's location, and is used to establish the spatial correspondence between the candidate defect and the parting line.

[0051] In one embodiment, after calculating the distance offset between the starting position of the candidate defect and the calibration reference point, the coordinates of the starting position of the candidate defect are corrected along the normal direction of the parting line based on the offset. The normal direction of the parting line refers to the vector perpendicular to the tangent direction of the parting line at the calibration reference point, and this vector points to one side of the adhesive strip application area. When correcting along the normal direction, the coordinates of the starting position of the candidate defect are translated towards the parting line direction by a translation distance equal to the projection component of the offset in the normal direction. This ensures that the corrected coordinates fall more accurately within the intersection area of ​​the parting line and the adhesive strip bonding surface, thus obtaining the calibrated coordinates of the candidate defect.

[0052] It should be noted that after the calibrated candidate defect coordinates are located in a specific area on the adhesive strip bonding surface, it is necessary to further determine whether this area belongs to the intersection area of ​​the adhesive strip bonding surface and the parting line. The direction vector of the adhesive strip laying path and the direction vector of the parting line at this coordinate are extracted, and the boundary range of the intersection area is determined based on the spatial relationship between the two direction vectors.

[0053] For example, the direction vector of the adhesive strip laying path extends along the length of the adhesive strip, and the direction vector of the parting line extends along the direction of the closed joint of the upper and lower molds. When the two direction vectors intersect at this coordinate, the coordinate is located within the intersection area. The boundary of the intersection area is defined by the width of the adhesive strip laying path and the influence range of the parting line, thus obtaining the precise spatial coordinates of each candidate defect within the intersection area. Further, the relationship between the flash and the adhesive strip path is compared and determined based on the precise spatial coordinates. The actual direction curve segment of the flash profile at this coordinate is read, the direction vector of the curve segment is extracted, and the laying direction vector of the adhesive strip path at this coordinate is read simultaneously. The flash direction vector represents the extension direction of the flash protrusion on the bonding surface, and the adhesive strip laying direction vector represents the travel direction of the adhesive strip laid along the edge of the shell. When calculating the angle between the two direction vectors, the cosine value of the angle is obtained by vector dot product operation, and then the angle value is obtained by inverse cosine operation. If the included angle is within a preset vertical angle threshold range, it is determined that the flash at that coordinate cuts into the adhesive strip path at a vertical or near-vertical angle, forming a linear channel that vertically crosses the adhesive strip path. The preset vertical angle threshold range is determined based on the critical conditions of the adhesive strip compression and rebound characteristics and the flash bridging effect, and is usually set to an angle range close to a right angle.

[0054] In one possible implementation, coordinates that meet the vertical crossing condition are marked as valid defect landing points, and the precise spatial coordinates of these valid defect landing points and their corresponding flash direction information are collected. The valid defect landing points are then sorted according to their distribution position on the adhesive strip bonding surface, based on the coordinate order along the adhesive strip laying direction.

[0055] Understandably, in areas with a high density of effective defect landing points, if the distance between adjacent landing points is less than a preset distance threshold, it indicates that these landing points point to the same leakage path caused by the burr, resulting in information redundancy. When removing redundant landing points where the distance between adjacent landing points is less than the preset distance threshold, the landing point with the largest step amplitude in each group of adjacent landing points is retained as a representative. The filtered effective defect landing points and their coordinate information are then incorporated into a refined defect coordinate list. This list records the defect location information after spatial calibration, vertical crossing verification, and redundancy removal, accurately marking the specific coordinates of the landing points on the adhesive strip bonding surface where linear leakage paths are formed due to the vertical crossing of burrs.

[0056] Another way to determine the refined defect coordinate list is to extract the calibration coordinates of each candidate defect in the area where the adhesive strip bonding surface and the parting line intersect, analyze the crossing situation of the flash direction and the adhesive strip sealing path at the coordinates, identify the defect landing points where the flash cuts through the adhesive strip path in a vertical manner and forms a continuous channel, and archive these landing points according to their positions on the sealing surface to obtain a refined defect coordinate list that has been spatially filtered and directionally verified.

[0057] The calibration coordinates of each candidate defect at the intersection of the adhesive strip bonding surface and the parting line are extracted. The direction curve segment of the flash profile and the laying direction of the adhesive strip sealing path at the calibration coordinates are read. The direction vector of the flash is sampled point by point along the flash direction curve segment, and the tangential direction vector of the adhesive strip sealing path is sampled at the same coordinate. The flash direction vector and the adhesive strip tangential direction vector are paired point by point to obtain the direction angle sequence at each sampling point. For each sampling point in the direction angle sequence, if the angle value is within the preset vertical angle threshold range, the sampling point is marked as a vertical cutting point. The number of consecutive vertical cutting points is counted. If the number of consecutive vertical cutting points exceeds the preset continuity threshold, it is determined that the flash at the calibration coordinate penetrates the adhesive strip sealing path in a vertical cutting manner, and the landing point of the candidate defect with vertical penetration characteristics is obtained. For candidate defect landing points with vertical penetration characteristics, the two ends of the burr's directional curve segment are checked to see if they extend to the two sides of the sealing path of the adhesive strip. If the start and end points of the burr's directional curve segment are located on opposite sides of the sealing path of the adhesive strip, it is determined that the burr forms a continuous channel penetrating the adhesive strip path at that coordinate. Defect landing points that meet the continuous channel condition are marked to obtain a set of defect landing points forming a continuous channel. The set of defect landing points forming a continuous channel is archived according to their spatial position on the sealing surface. Based on the calibration coordinates of each defect landing point arranged along the adhesive strip laying direction, and combined with the burr's directional direction and vertical cutting angle information, a refined list of defect coordinates after spatial screening and directional verification is obtained.

[0058] During the speaker enclosure sealing performance testing, the calibration coordinates of each candidate defect at the intersection of the adhesive strip bonding surface and the parting line are derived from the spatial calibration results of the initial positioning coordinate set and the parting line path coordinates. After extracting these calibration coordinates, the direction curve segment of the flash profile and the laying direction of the adhesive strip sealing path at that coordinate are read. The direction vector is sampled point by point along the flash direction curve segment, and the sampling interval is determined according to the curvature change of the flash profile.

[0059] Specifically, several sampling points are set along the burr curve segment, and the tangent direction of the burr curve is calculated at each sampling point as the direction vector of that point. Simultaneously, the tangent direction vector is sampled at the calibration coordinates along the sealing path of the adhesive strip; this direction vector represents the laying direction of the adhesive strip at that location. After pairing the burr direction vector with the adhesive strip tangent direction vector point by point, the angle between each pair of vectors is calculated, forming a sequence of direction angles covering the entire burr curve segment.

[0060] In one embodiment, a vertical cut-in determination is performed for each sampling point in the directional angle sequence. The preset vertical angle threshold range is determined based on the compression and rebound characteristics of the rubber material of the adhesive strip. When the angle between the flash direction and the adhesive strip laying direction is close to a right angle, the flash protrusion will create a local gap in the width direction of the adhesive strip. If the angle value of a sampling point falls within this threshold range, the sampling point is marked as a vertical cut-in point. The marking status of each sampling point is scanned sequentially along the directional angle sequence, and the number of consecutively occurring vertical cut-in points is counted.

[0061] It should be noted that the preset continuity threshold is used to determine whether the burr maintains a vertical cut-in state within a sufficient length range. If the number of consecutive vertical cut-in points exceeds this threshold, it indicates that the burr cuts into the sealing path of the adhesive strip at a stable vertical angle at the calibration coordinate, possessing the geometric conditions to form an overhead channel. Thus, it is determined that the burr at the calibration coordinate penetrates the sealing path of the adhesive strip in a vertical cut-in manner, obtaining a candidate defect landing point with vertical penetration characteristics. Further, a continuous channel penetration detection is performed on the candidate defect landing points with vertical penetration characteristics. The two ends of the burr's direction curve segment are checked to see if they extend to the two side boundaries of the sealing path of the adhesive strip, where the two side boundaries refer to the left and right edge lines of the adhesive strip along its width direction. The positional relationship between the coordinates of the starting point of the burr's direction curve segment and the left edge line of the adhesive strip is read, and the positional relationship between the coordinates of the ending point of the burr's direction curve segment and the right edge line of the adhesive strip is read. If the starting point is located outside the left edge line of the adhesive strip and the ending point is located outside the right edge line of the adhesive strip, or vice versa, it indicates that the burr's direction curve segment completely spans the width direction of the adhesive strip.

[0062] For example, when the flash cuts into the rubber strip from one edge at a vertical angle and exits from the other edge, the flash protrusion forms a continuous overhead ridge line inside the rubber strip. This ridge line connects the gaps between the upper surface of the rubber strip and the lower shell surface into a passage. After determining that the flash constitutes a continuous channel through the rubber strip at this coordinate, the defect points that meet the conditions for a continuous channel are marked, resulting in a set of defect points that constitute a continuous channel.

[0063] In one possible implementation, the set of defect landing points constituting the continuous channel is archived according to their spatial positions on the sealing surface. Based on the order of the calibration coordinates of each defect landing point along the adhesive strip laying direction, the relative position of each landing point on the adhesive strip path is recorded.

[0064] It is understood that the refined defect coordinate list not only includes the spatial coordinate information of each defect landing point, but also the direction of the flash and the vertical cutting angle at that landing point. The flash direction is used to identify the extension direction of the leakage channel, and the vertical cutting angle is used to quantify the degree of intersection between the flash and the adhesive strip path. After integrating the spatial screening results and the direction verification results, a refined defect coordinate list after spatial screening and direction verification is formed. This list accurately marks the specific location and characteristic information of the continuous leakage channel formed by the vertical penetration of the flash on the adhesive strip bonding surface.

[0065] Step S105: Read the combination features of compression mutation gradient and step amplitude at the corresponding positions one by one according to the refined defect coordinate list, group similar defect types, and determine the leakage severity index of each defect type.

[0066] The compression gradient and step amplitude at each defect location are read sequentially from the refined defect coordinate list. The compression gradient represents the rate of change of compression along the width of the rubber strip after being lifted by the flash at that location. The step amplitude represents the numerical drop in pressure at that location during the pressure test. The compression gradient and step amplitude of each defect location are paired and combined to obtain a combined feature vector for each defect location. For the combined feature vectors of each defect location, a hierarchical clustering method is used to measure similarity based on the Euclidean distance between the combined feature vectors. Defect locations with a combined feature vector distance less than a preset clustering threshold are grouped into the same group. After traversing all defect locations and completing the grouping, several defect groups are obtained. The defect locations within each defect group have similar compression gradient and step amplitude characteristics. Each defect group is marked as the corresponding defect type. For each defect type, the average step amplitude A and the average compressed mutation gradient G of all defect points within that type are read, where A is the average step amplitude and G is the average compressed mutation gradient. A and G are then summed using weighted coefficients. The weighting coefficients are determined empirically as follows: the weight w1 for the step amplitude is 0.6 and the weight w2 for the compressed mutation gradient is 0.4. The calculation formula is S = w1 * A + w2 * G, which yields the leakage severity index S for that defect type.

[0067] For example, if A=5.2 and G=3.8, then S=0.6*5.2+0.4*3.8.

[0068] During the speaker enclosure sealing performance testing, the compression gradient and step amplitude at each defect location in the refined defect coordinate list are derived from the processing results of the diaphragm data collected in the previous stage of pressure distribution. The compression gradient characterizes the rate of change of the compression amount along the width direction of the adhesive strip after it is lifted by the flash at that location. The larger the gradient value, the more significant the disturbance to the uniformity of the adhesive strip at that location.

[0069] Specifically, the step amplitude characterizes the numerical difference in pressure drop at that location during the pressurization test, reflecting the severity of gas escape after the formation of the leak channel. By pairing and combining the compression gradient and step amplitude at each defect location, a combined feature vector for that defect location is formed. This vector contains two components, corresponding to the compression gradient value and the step amplitude value, respectively.

[0070] In one embodiment, a hierarchical clustering method is used to group the combined feature vectors of each defect landing point. The hierarchical clustering method progressively merges defect landing points with high similarity in a bottom-up manner, using the Euclidean distance between the combined feature vectors as the similarity measure. The Euclidean distance is calculated as the square root of the sum of the squares of the differences between the compressed abrupt gradient component and the step amplitude component of two combined feature vectors; the smaller the distance, the more similar the features of the two defect landing points. Defect landing points whose combined feature vector distance is less than a preset clustering threshold are grouped into the same group. After traversing all defect landing points and completing hierarchical merging, several defect groups are obtained. The defect landing points within each defect group have similar compressed abrupt gradient and step amplitude features.

[0071] It should be noted that each defect group is labeled with a corresponding defect type, and different defect types represent different combinations of leakage characteristics. Further, a leakage severity index is calculated for each defect type. The average value of the step amplitude at all defect points within that type is calculated, and the average value of the compression gradient is also calculated. The average values ​​of the step amplitude and compression gradient are then weighted and summed according to a preset weighting coefficient, which is determined based on the relative importance of the step amplitude and compression gradient to airtightness. The weighted sum yields the leakage severity index for that defect type. A higher index value indicates a higher risk of seal failure caused by that type of defect, facilitating subsequent risk level classification of different defect types.

[0072] Step S106: Archive the leakage severity index according to the three thresholds of slight, moderate and severe in the airtightness classification standard according to the numerical range, mark the defects in the moderate and above range as high-risk leakage channels, and generate the airtightness classification result of the sealing surface.

[0073] The leakage severity index for each defect type is obtained. The threshold ranges corresponding to the three levels (minor, moderate, and severe) in the preset airtightness grading standard are read. The leakage severity index for each defect type is compared one by one with the three threshold ranges. If the index value falls into the minor range, it is classified as minor; if it falls into the moderate range, it is classified as moderate; and if it falls into the severe range, it is classified as severe. This yields an airtightness grade label for each defect type. For each defect type's airtightness grade label, defect types with airtightness grade labels of moderate or severe are selected. All defect points corresponding to this type are marked as high-risk leakage channels. The airtightness grade labels and high-risk marking statuses for each defect type are combined to generate the airtightness grading result for the sealing surface.

[0074] During the speaker enclosure sealing performance testing, the leakage severity index for each defect type is derived from the weighted calculation results of the compression gradient and step amplitude. After obtaining the leakage severity index, the threshold ranges corresponding to the three levels of slight, moderate, and severe in the preset airtightness grading standard are read. These threshold ranges are determined based on the speaker enclosure sealing performance requirements and historical test data.

[0075] Specifically, the leakage severity index for each defect type is compared with the three threshold ranges one by one. If the index value falls into the slight range, it is classified as slight level; if it falls into the medium range, it is classified as medium level; and if it falls into the severe range, it is classified as severe level, thus obtaining the airtightness level label for each defect type.

[0076] In one embodiment, when performing high-risk screening for the airtightness level labels of each defect type, defect types with airtightness level labels classified as medium or severe are selected, and all defect landing points corresponding to that type are marked as high-risk leakage channels. After aggregating the airtightness level labels and high-risk marking statuses for each defect type, a sealing surface airtightness classification result is formed, which reflects the risk level distribution of various defects on the sealing surface.

[0077] Step S107: The airtightness grading results of the sealing surface are superimposed with the three-dimensional scanning image of the adhesive strip bonding surface. The coordinates of each defect are mapped to the corresponding position on the adhesive strip bonding surface according to the grading results to obtain the location result of the sealing defect.

[0078] The calibration coordinates of each defect landing point are obtained from the airtightness classification results. These coordinates are obtained by least squares fitting after defect detection algorithms such as Canny edge detection. The spatial coordinate reference and scale information of the 3D scan image are read, and an image registration reference system is established based on the spatial coordinate reference. The calibration coordinates are transformed into the image registration reference system to obtain the mapped coordinates of each defect landing point in the 3D scan image. For the mapped coordinates of each defect landing point in the 3D scan image, the airtightness level label and high-risk marking status corresponding to the defect landing point are read, and superimposed annotations are made at the corresponding mapped coordinate positions in the 3D scan image. Different annotation symbols are used to distinguish between three levels of airtightness: minor, medium, and severe. Defect landing points in high-risk leakage channels are highlighted, resulting in an image of the adhesive strip bonding surface after superimposed annotation. Based on the superimposed and annotated image of the adhesive strip bonding surface, the contour extraction function of image processing algorithms such as OpenCV is used to further verify and collect the mapping coordinates of each defect landing point, the airtightness level label, the high-risk marking status and the direction information of the burr. This direction information is obtained from the defect edge morphology analysis as a vector along the adhesive strip. The vector is arranged according to the order of the defect landing points in the adhesive strip laying direction to form a complete record containing the spatial location and airtightness classification information of the defect, thus obtaining the final sealing defect orientation location result.

[0079] In the process of testing the sealing performance of speaker housings, the overlay processing of the airtightness grading results of the sealing surface and the three-dimensional scanning image of the adhesive strip bonding surface is a key step in realizing the visualization and location of defects. After obtaining the airtightness grading results of the sealing surface, the three-dimensional scanning image of the adhesive strip bonding surface is read simultaneously. This image records the complete surface morphology information of the area where the adhesive strip is applied.

[0080] Specifically, when reading the spatial coordinate reference and scale information of a 3D scanned image, the spatial coordinate reference refers to the origin position and coordinate axis direction set by the scanner when acquiring the image, and the scale information refers to the conversion relationship between image pixels and actual physical dimensions. An image registration reference system is established based on the spatial coordinate reference. This reference system unifies the defect point calibration coordinates in the airtightness grading results with the pixel coordinates of the 3D scanned image under the same spatial measurement system.

[0081] In one embodiment, when converting the calibration coordinates of each defect landing point in the airtightness grading result to the image registration reference system, the physical distance value of the calibration coordinates is converted into the pixel position index in the three-dimensional scan image according to the scale information, so as to obtain the mapping coordinates of each defect landing point in the three-dimensional scan image.

[0082] It should be noted that when overlaying and annotating the mapped coordinates of each defect landing point in the 3D scan image, the corresponding airtightness level label and high-risk marking status are read. Overlay annotations are performed at the corresponding mapped coordinate positions in the 3D scan image, using different shapes or colors of annotation symbols to distinguish between minor, moderate, and severe airtightness levels: a circle for minor, a triangle for moderate, and a square for severe. High-risk leakage channels are marked with highlighted borders to enhance their visual visibility in the image, resulting in an image of the adhesive strip bonding surface after overlay annotation. Further, based on the image of the adhesive strip bonding surface after overlay annotation, the mapped coordinates, airtightness level labels, high-risk marking status, and burr direction information of each defect landing point are collected. After organizing the defect landing points according to their arrangement in the adhesive strip laying direction, a complete record containing the spatial location and airtightness classification information of the defects is formed, resulting in the final sealing defect location result. This location result presents the specific location and risk level of each defect on the adhesive strip bonding surface in a combination of visualized images and structured data.

[0083] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the invention. Those skilled in the art will understand that implementing all or part of the above embodiments and making equivalent changes in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A method for evaluating the sealing degree of a speaker enclosure, characterized in that, The method includes: Acquire three-dimensional scanning images of the adhesive strip bonding surface, coordinates of the parting line path, and pressure readings of the diaphragm during the pressure test in the sealed cavity. Extract the flash profile based on the three-dimensional scanning images of the adhesive strip bonding surface and calculate the intersection angle between the flash and the adhesive strip direction. Based on the intersection angle between the flash and the direction of the rubber strip, potential linear channel risk areas are identified and marked. Their spatial range is mapped to the corresponding coordinates of the pressure readings. Local abrupt changes in the pressure readings within this range are analyzed and converted into compression abrupt change gradients. The pressure readings of the diaphragm during the pressurization test in the sealed cavity are fitted along the time axis to generate an overall leakage rate curve of the sealing surface. The time and magnitude of the step in the curve are identified, and the spatial coordinates of the step point and the potential linear channel risk area are associated to obtain the initial location coordinate set of directional leakage. Based on the initial positioning coordinate set and the parting line path coordinates, the spatial position of each candidate defect is recalibrated to obtain its precise spatial coordinates in the area where the adhesive strip and the parting line intersect. The relationship between the actual direction of the flash and the adhesive strip path is compared at this coordinate. The coordinate landing points of the linear channel that forms a vertical passage through the adhesive strip path are screened and added to the refined defect coordinate list. Read the combined characteristics of compression mutation gradient and step amplitude at the corresponding position one by one according to the refined defect coordinate list, group similar defect types, and determine the leakage severity index of each defect type. The severity of leakage indicators are filed one by one according to the three thresholds of slight, moderate and severe in the airtightness classification standard according to the numerical range. Defects in the moderate and above ranges are marked as high-risk leakage channels, and the airtightness classification results of the sealing surface are generated. The airtightness classification results are superimposed with the three-dimensional scan image to obtain the location result of the sealing defect.

2. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The step of extracting the burr contour from the 3D scan image of the adhesive strip bonding surface and calculating the intersection angle between the burr and the adhesive strip direction includes: Surface point cloud data is extracted from the three-dimensional scanned image, and cross-sectional slices are generated along the boundary of the adhesive strip application area. Extract the surface height value of each slice, and mark the area where the height difference between adjacent slices exceeds the preset benchmark as a convex candidate area; By vectorizing and connecting the edges of the candidate protrusion areas, the spatial orientation curve of the flash profile is obtained; Generate a reference baseline based on the path coordinates of the parting line, and read the laying direction of the adhesive strip; The burr outline is projected onto the bonding surface to form a projection line segment. The angle between the projection line segment and the direction of the adhesive strip is measured to obtain the intersection angle value.

3. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The process of identifying and marking potential linear channel risk areas based on the intersection angle between the flash and the adhesive strip direction, mapping their spatial range to the corresponding coordinates of the pressure readings, analyzing local abrupt changes in pressure readings within this range, and converting them into compression gradient changes includes: Filter out the projected line segments whose intersection angle exceeds a preset vertical threshold, and mark the covered area as a potential linear channel risk zone; Extract the boundary coordinate range of the risk zone and establish a mapping relationship between the coordinates of the adhesive strip bonding surface and the coordinates of the pressure distribution data grid; The risk zone boundary coordinates are converted into grid row and column index intervals to obtain the coordinate mapping result; For the grid interval defined by the coordinate mapping result, read the pressure value of each grid node; The pressure distribution is scanned along the width of the adhesive strip to identify the locations where the pressure difference between adjacent grid nodes exceeds the preset mutation threshold. The locations of local mutations are determined, and the ratio of pressure difference to grid spacing is calculated to obtain the compression mutation gradient.

4. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The pressure readings of the pressure distribution acquisition diaphragm during the pressurization test in the sealed cavity are fitted along the time axis to generate an overall leakage rate curve of the sealing surface. The timing and magnitude of the step in the curve are identified, including: For each grid node, the pressure change trajectory is plotted, and the decay rate is extracted using a fitting method. Summarize the decay rates of each node to generate an overall leakage rate curve for the sealing surface; Scan along the time axis of the leakage rate curve to identify locations where the pressure difference between adjacent sampling points exceeds a preset change threshold, and mark them as step moments; Read the pressure drop before and after the step moment and record the step amplitude.

5. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The spatial coordinates of the associated step points and the potential linear channel risk area are used to obtain the initial location coordinate set of the directional leakage, including: sorting and filtering risk-associated step points by step amplitude, mapping grid coordinates to the sealing surface, and summarizing to obtain the initial location coordinate set of the directional leakage.

6. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, Based on the initial positioning coordinate set and the parting line path coordinates, the spatial position of each candidate defect is recalibrated to obtain its precise spatial coordinates in the area where the adhesive strip bonding surface intersects with the parting line. At these coordinates, the relationship between the actual direction of the flash and the adhesive strip path is compared. The coordinate points that form a linear channel perpendicular to the adhesive strip path are then filtered and added to the refined defect coordinate list, including: The starting positions of each candidate defect in the initial positioning coordinate set are obtained. The nearest node is found along the path coordinates of the parting line as a calibration reference point. The precise spatial coordinates are obtained by correcting the offset along the normal direction. Extract the vectors of the actual direction of the burr and the path of the adhesive strip, calculate the angle between the actual direction of the burr and the path of the adhesive strip, and determine that a vertical crossing channel is formed when the angle is within the vertical threshold range and mark the effective defect landing point. Based on the distribution location, redundant items that are too close to each other are sorted and redundant items are removed. The effective defect landing points are recorded to obtain a list of defect coordinates.

7. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The coordinate points of the linear channels that form the vertical path through the adhesive strip are filtered and imported into the refined list of defect coordinates, including: Extract the calibration coordinates of the area where the adhesive strip mating surface intersects with the parting line, read the flash outline and the sealing laying direction, and obtain the included angle by sampling the direction vector pairing; If the included angle is within the threshold, the coordinates are marked as vertical entry points. When the number of consecutive vertical entry points exceeds the continuity threshold, the coordinates are determined to be candidate defect landing points. Arrange and archive the defects according to their spatial location on the sealing surface, filter out the candidate defect locations that meet the preset conditions, and obtain the refined defect coordinate list.

8. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The process involves sequentially reading the combined features of compression abrupt change gradient and step amplitude at corresponding locations according to the refined defect coordinate list, grouping similar defect types, and determining the leakage severity index for each defect type, including: Read the compression abrupt gradient and step amplitude of each point in the list of refining defect coordinates, and pair them to form a combined feature vector; Clustering is used to measure similarity by feature vector distance. Points with a distance less than a preset threshold are grouped into the same group to obtain defect type groups. For each defect type group, the mean values ​​of step amplitude and compression mutation gradient within each group are calculated. The leakage severity index for each defect type is obtained by weighting and summing the mean values ​​of the step amplitude and compression mutation gradient within each group according to preset weights.

9. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The process involves classifying leakage severity indicators according to numerical ranges and comparing them to the three threshold levels of minor, moderate, and severe in the airtightness grading standard. Defects in the moderate and above ranges are marked as high-risk leakage channels, generating an airtightness grading result for the sealing surface, including: The severity index of the leak is obtained and compared with the three threshold ranges of the airtightness classification standard. If the leak falls into the corresponding range, it is classified as slight, moderate or severe, and an airtightness level label is obtained. The airtightness rating labels are selected as medium or severe, and all corresponding defect locations are marked as high-risk leakage channels. The airtightness rating labels and marking status are combined to obtain the airtightness classification result of the sealing surface.

10. The method for evaluating the sealing degree of a speaker housing according to claim 1, characterized in that, The process of overlaying the airtightness grading results of the sealing surface with the three-dimensional scan image of the adhesive strip bonding surface, and mapping the coordinates of each defect to the corresponding position on the adhesive strip bonding surface according to the grading results, yields the location results of the sealing defects, including: A reference system is established by reading the spatial coordinates of the three-dimensional scanned image of the adhesive strip bonding surface, and the coordinates of the defect landing point are transformed to obtain the mapped coordinates. The airtightness level label and high-risk status are superimposed at the corresponding mapping coordinates. Different symbols are used to distinguish the airtightness levels corresponding to the three thresholds. High-risk leakage channels are marked with highlighted labels to obtain the superimposed image. The spatial location of the defect and the airtightness information are arranged in order to obtain the location result of the sealing defect.

Images