Electronic chip package calibration method and system based on visual positioning

Abstract

This invention relates to the field of electronic chip image processing technology, and discloses a method and system for electronic chip packaging calibration based on visual positioning. The method includes: acquiring a reference chip packaging image based on the chip to be calibrated; using a visual positioning unit to acquire an image of the chip to be calibrated, obtaining a chip packaging image; identifying chip feature regions in the chip packaging image to obtain a group of effective feature regions; performing packaging deviation calibration based on the feature regions to be matched and the extracted effective feature regions, obtaining a target deviation correction vector; averaging all target deviation correction vectors to obtain an average deviation correction vector; and generating a chip packaging calibration report based on the average deviation correction vector. This invention can improve the computational efficiency and overall stability of electronic chip packaging calibration, and reduce sensitivity to irrelevant image noise.

Description

Technical Field

[0001] This invention relates to the field of electronic chip image processing technology, and in particular to an electronic chip packaging calibration method and system based on visual positioning. Background Technology

[0002] In modern electronics manufacturing, the precision of chip packaging directly determines the performance and reliability of the final product. With the continuous improvement of chip integration and the continuous miniaturization of package size, micron or even nanometer-level pose deviations during the packaging process can lead to serious problems such as electrical connection failure and reduced heat dissipation performance. Therefore, achieving high-precision and high-reliability chip packaging calibration is a key technical link to ensure the quality of advanced electronic products and improve production yield, and has important industrial application value.

[0003] Traditional techniques typically involve directly comparing pixels or matching features between the captured high-resolution image of the chip to be calibrated and the reference image to calculate the deviation. This approach involves a large-scale exhaustive search and matching at a fine scale (high resolution), which is computationally complex and time-consuming. Furthermore, this method is sensitive to image noise and small, statistically insignificant local deviations, which can easily lead to over-adjustment of irrelevant details and may introduce unstable calibration commands. Summary of the Invention

[0004] This invention provides a visual positioning-based electronic chip packaging calibration method and a computer-readable storage medium. Its main purpose is to improve the computational efficiency and overall stability of electronic chip packaging calibration, and reduce the sensitivity to irrelevant image noise.

[0005] To achieve the above objectives, the present invention provides a vision-based positioning-based electronic chip packaging calibration method, comprising:

[0006] Receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0007] The pre-built visual positioning unit is used to acquire images of the chip to be calibrated, thereby obtaining the packaged image of the chip to be calibrated;

[0008] The chip feature region is identified from the chip package image to be calibrated to obtain a group of effective feature regions;

[0009] Effective feature regions are extracted sequentially from the effective feature region group, and the matching feature regions corresponding to the extracted effective feature regions are identified from the chip feature region group in the reference chip package image.

[0010] Encapsulation deviation calibration is performed based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector;

[0011] The target deviation correction vectors corresponding to each effective feature region are summarized to obtain the target deviation correction vector group. The target deviation correction vector group is then averaged to obtain the average deviation correction vector. Based on the average deviation correction vector, a chip packaging calibration report is generated, thus completing the vision-based positioning-based electronic chip packaging calibration.

[0012] Optionally, obtaining the reference chip package image based on the chip to be calibrated includes:

[0013] Obtain chips of the same model based on the chip to be calibrated;

[0014] A standard chip packaging base is obtained by precisely packaging chips of the same model using a preset standard packaging base. The standard chip packaging base contains chips of the same model.

[0015] The chip is photographed using a visual positioning unit on a standard chip packaging base to obtain an initial chip packaging image;

[0016] Chips of the same model are divided into feature regions to obtain chip feature region groups, where each chip feature region group includes multiple chip feature regions.

[0017] The initial chip package image is marked with regions based on the chip feature region group to obtain the reference chip package image.

[0018] Optionally, the step of identifying chip feature regions in the chip package image to be calibrated to obtain a group of effective feature regions includes:

[0019] The image of the chip package to be calibrated is converted to grayscale to obtain a grayscale image of the chip to be calibrated. Then, grayscale threshold segmentation is performed on the grayscale image of the chip to be calibrated to obtain a segmented image of the chip to be calibrated, wherein the segmented image of the chip to be calibrated includes multiple pixels to be calibrated.

[0020] Chip contour extraction is performed on the segmented image of the chip to be calibrated to obtain the chip contour region;

[0021] The current center pixel group is identified based on the chip feature region group and the chip contour region. The current center pixel group includes multiple current center pixels, and each current center pixel corresponds one-to-one with a chip feature region.

[0022] Based on the current group of center pixels, perform region growing to obtain a group of effective feature regions.

[0023] Optionally, the step of identifying the current center pixel group based on the chip feature region group and the chip contour region includes:

[0024] The standard contour region is extracted based on the reference chip packaging image, and a coordinate system is constructed for the standard contour region to obtain the standard contour coordinate system.

[0025] The chip contour region is scaled and transformed based on the standard contour region to obtain the target contour region, wherein the coordinate system of the target contour region is the same as the coordinate system of the standard contour region.

[0026] Extract chip feature regions sequentially from the chip feature region group and identify the standard region center of the extracted chip feature regions;

[0027] Determine the standard center coordinates of the standard region center based on the standard contour coordinate system;

[0028] Determine the current center pixel point in the target contour region that corresponds to the standard center coordinates;

[0029] The current center pixel points corresponding to each chip feature region are summarized to obtain the current center pixel point group.

[0030] Optionally, the step of performing region growing based on the current group of center pixels to obtain a group of effective feature regions includes:

[0031] Extract the current center pixel from the current center pixel group in sequence, and record the extracted current center pixel as the starting pixel;

[0032] Get the starting pixel value of the starting pixel;

[0033] In the segmented image of the chip to be calibrated, identify the set of neighboring pixels corresponding to the starting pixel, where the set of neighboring pixels includes multiple neighboring pixels.

[0034] Obtain the set of neighboring pixel values ​​of the set of neighboring pixels, where the neighboring pixel values ​​in the set of neighboring pixel values ​​correspond one-to-one with the neighboring pixels;

[0035] Based on the adjacent pixel value set and the starting pixel value, the same pixel value is judged to obtain the pixel judgment result, where the pixel judgment result is either that there are the same pixel values ​​or that there are no the same pixel values;

[0036] If the pixel discrimination result is that there are identical pixel values, then the pixel group in the adjacent pixel set is identified based on the adjacent pixel value set and the starting pixel value.

[0037] The pixel group within the domain is added to the pre-constructed original feature region pixel set to obtain the target feature region pixel set;

[0038] The target feature region pixel set and the domain pixel group are respectively used as the original feature region pixel set and the current center pixel group, and the step of extracting the current center pixel in the current center pixel group is returned until the pixel discrimination result is that there are no identical pixel values.

[0039] If the pixel discrimination result is that there are no identical pixel values, then the original feature region pixel set is recorded as the effective feature region pixel set;

[0040] The effective feature region is divided according to the set of pixels in the effective feature region;

[0041] The effective feature regions corresponding to each current center pixel are summarized to obtain the effective feature region group.

[0042] Optionally, the step of performing encapsulation deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector includes:

[0043] Based on the extracted effective feature regions, the packaging region is selected in the grayscale image of the chip to be calibrated to obtain the effective packaging region. Based on the feature regions to be matched, the reference packaging region is selected in the reference chip packaging image.

[0044] Pyramid downsampling is performed on the effective encapsulation region to obtain an effective pyramid image set, which includes multiple effective pyramid images;

[0045] Obtain a reference pyramid image set for the reference encapsulation region, wherein the reference pyramid image set includes multiple reference pyramid images, and each reference pyramid image corresponds one-to-one with a valid pyramid image;

[0046] The effective pyramid images are extracted sequentially from the effective pyramid image set, and the extracted effective pyramid images are recorded as the top effective pyramid images. The top effective pyramid images are then transformed using a preset initial deviation correction vector to obtain the transformed top pyramid images.

[0047] Identify the reference pyramid top-level image corresponding to the valid pyramid top-level image in the reference pyramid image set;

[0048] The target deviation correction vector is constructed based on the top-level image of the baseline pyramid and the top-level image of the transformed pyramid.

[0049] Optionally, constructing the target deviation correction vector based on the reference pyramid top-level image and the transformed pyramid top-level image includes:

[0050] The top image of the transform pyramid is encapsulated with the top image of the base pyramid to identify the deviation, and the deviation identification result is obtained, which indicates whether the deviation exists or not.

[0051] If the deviation identification result has a deviation, then obtain the encapsulated deviation correction vector;

[0052] The encapsulated deviation correction vector is used as the initial deviation correction vector, and the step of sequentially extracting valid pyramid images in the valid pyramid image set is returned until the deviation identification result is that there is no deviation.

[0053] If the deviation identification result is that there is no deviation, then the initial deviation correction vector is recorded as the target deviation correction vector.

[0054] Optionally, the step of using the top-level image of the reference pyramid to perform encapsulation deviation identification on the top-level image of the transformed pyramid to obtain the deviation identification result includes:

[0055] The set of transformed pixels in the top layer of the transform pyramid image is identified, wherein the set of transformed pixels includes multiple transformed pixels;

[0056] Transformed pixels are extracted sequentially from the set of transformed pixels, and the neighborhood of transformed pixels is identified in the top layer image of the transformation pyramid based on the extracted transformed pixels and the preset matching template size.

[0057] Region matching is performed on the neighborhood of transformed pixels based on the top-level image of the baseline pyramid to obtain the region matching degree.

[0058] Summarize the region matching degree corresponding to each transformed pixel to obtain the region matching degree set;

[0059] Identify the maximum regional deviation value based on the regional matching degree set, where the maximum regional deviation value is the minimum value in the regional matching degree set;

[0060] Obtain the target transformation region and target reference region corresponding to the maximum regional deviation value;

[0061] The target transformation region is solved by using the target reference region to obtain the region correction data, which includes: the region correction horizontal axis displacement, the region correction vertical axis displacement, and the region clockwise rotation.

[0062] The regional correction amount is calculated based on the regional correction data. If the regional correction amount is greater than the preset standard correction amount, the existence of deviation is recorded as the deviation identification result; otherwise, the absence of deviation is recorded as the deviation identification result.

[0063] Optionally, the step of performing region matching on the transformed pixel neighborhood based on the top-level image of the benchmark pyramid to obtain the region matching degree includes:

[0064] Feature extraction is performed on the neighborhood of the transformed pixels to obtain the features of the transformed region.

[0065] Based on the matching template size, the top layer image of the baseline pyramid is traversed to obtain the traversed baseline image region.

[0066] Obtain the traversal region features of the reference image region, and calculate the region matching degree based on the transformed region features and the traversal region features.

[0067] Return to the step of traversing the top-level image of the baseline pyramid based on the matching template size, until the traversal is complete;

[0068] Summarize the regional matching scores to obtain the regional matching score set, identify the maximum matching score in the regional matching score set, and record the maximum matching score as the regional matching score.

[0069] To achieve the above objectives, the present invention also provides a vision-based positioning-based electronic chip packaging calibration system, comprising:

[0070] The packaging instruction receiving module is used to receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0071] The feature region recognition module is used to acquire images of the chip to be calibrated using a pre-built visual positioning unit, obtain a packaged image of the chip to be calibrated, and perform chip feature region recognition on the packaged image of the chip to be calibrated to obtain a group of effective feature regions.

[0072] The standard region matching module is used to extract effective feature regions sequentially from the effective feature region group and identify the feature region to be matched corresponding to the extracted effective feature region in the chip feature region group in the reference chip package image.

[0073] The calibration report generation module is used to perform packaging deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector. It summarizes the target deviation correction vectors corresponding to each effective feature region to obtain the target deviation correction vector group, and generates a chip packaging calibration report based on the average deviation correction vector.

[0074] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0075] Memory, storing at least one instruction;

[0076] The processor executes the instructions stored in the memory to implement the vision-based positioning electronic chip packaging calibration method described above.

[0077] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned vision-based positioning-based electronic chip packaging calibration method.

[0078] To address the problems described in the background section, this invention first obtains a reference chip package image based on the chip to be calibrated. This step involves precisely packaging a chip of the same model with a standard packaging base and capturing the image, constructing a physically accurate reference image containing known feature location information, rather than relying on theoretical drawings or CAD models. This provides a high-fidelity reference standard for subsequent visual comparison, avoiding systematic errors introduced by differences between models and physical objects. Furthermore, this solution performs chip feature region identification on the chip package image to be calibrated, obtaining a group of effective feature regions. This step, through image processing, contour extraction, coordinate system mapping, and a region growing algorithm, can automatically and accurately segment and identify multiple feature regions corresponding to the reference image from the chip image to be calibrated. This method ensures the reliability of feature recognition, providing a foundation for subsequent precise... The accurate alignment lays the foundation. Finally, based on the feature regions to be matched and the extracted effective feature regions, packaging deviation calibration is performed to obtain the target deviation correction vector. This step adopts a coarse-to-fine pyramid hierarchical matching and iterative correction strategy. Compared with the traditional approach of directly performing global search matching on the original high-resolution image, this method first performs fast matching and preliminary deviation correction at the low-resolution image layer, which has a small computational load and can quickly eliminate large offsets. Then, the preliminary result is used as the initial value for iterative fine adjustment at the higher-resolution image layer. This strategy avoids large-scale and time-consuming exhaustive searches at a fine scale, reducing the overall computation time while ensuring the final calibration accuracy. At the same time, the introduced early termination mechanism (stopping when there is no deviation in a certain layer) further optimizes computational resources and avoids unnecessary fine calculations. Therefore, this invention can improve the computational efficiency and overall stability of electronic chip packaging calibration and reduce sensitivity to irrelevant image noise. Attached Figure Description

[0079] Figure 1 This is a flowchart illustrating a vision-based electronic chip packaging calibration method according to an embodiment of the present invention.

[0080] Figure 2 A functional block diagram of a vision-based electronic chip packaging and calibration system provided in an embodiment of the present invention;

[0081] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the vision-based positioning electronic chip packaging calibration method according to an embodiment of the present invention.

[0082] Explanation of reference numerals in the attached figures:

[0083] 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0084] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0085] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0086] This application provides a visual positioning-based electronic chip packaging calibration method. The execution entity of the visual positioning-based electronic chip packaging calibration method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the visual positioning-based electronic chip packaging calibration method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0087] Reference Figure 1 The diagram shown is a flowchart illustrating a vision-based electronic chip packaging calibration method according to an embodiment of the present invention. In this embodiment, the vision-based electronic chip packaging calibration method includes:

[0088] S1. Receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0089] It is clear that the electronic chip packaging instruction refers to a manually initiated instruction to calibrate the packaging of a specific chip. The chip to be calibrated refers to the specific chip included in the electronic chip packaging instruction. The reference chip packaging image refers to an image captured by a vision positioning unit after the chip to be calibrated has been precisely packaged.

[0090] Specifically, obtaining the reference chip package image based on the chip to be calibrated includes:

[0091] Obtain chips of the same model based on the chip to be calibrated;

[0092] A standard chip packaging base is obtained by precisely packaging chips of the same model using a preset standard packaging base. The standard chip packaging base contains chips of the same model.

[0093] The chip is photographed using a visual positioning unit on a standard chip packaging base to obtain an initial chip packaging image;

[0094] Chips of the same model are divided into feature regions to obtain chip feature region groups, where each chip feature region group includes multiple chip feature regions.

[0095] The initial chip package image is marked with regions based on the chip feature region group to obtain the reference chip package image.

[0096] Understandably, the "same model chip" refers to a chip with the same model number as the chip to be calibrated. The "standard packaging base" refers to a carrier used to support the same model chip, such as a standardized test socket for BGA packaging. The "standard chip packaging base" refers to the assembly formed by precisely placing and fixing the same model chip onto the standard packaging base, used for capturing reference images. The above-mentioned precise packaging of the same model chip using a preset standard packaging base refers to: picking up the same model chip using a robotic arm or precision pick-and-place device, precisely aligning it with preset positioning marks (such as guide edges and positioning holes) on the standard packaging base, then placing and fixing it, and finally having relevant operators inspect the fixed same model chip and standard packaging base using a high-powered microscope or laser displacement sensor, etc., to ensure that there is no movement or rotation deviation during the fixing. The assembly of the fixed standard packaging base and the same model chip is then recorded as the standard chip packaging base.

[0097] Furthermore, the initial chip package image refers to a top-view image of a standard chip package base captured by a visual positioning unit. The chip feature region group refers to a set of multiple chip feature regions, where each chip feature region is a manually defined local area that represents the geometric or textural features of a chip of the same model. Examples include: the area adjacent to the apex of a chip of the same model (e.g., a circular area with a radius of several pixels centered at the apex), and specific surface texture patterns (e.g., trademarks, serial numbers). The process of marking the initial chip package image according to the chip feature region group refers to clearly identifying the specific location and extent of each chip feature region on the initial chip package image by drawing bounding boxes, polygons, or adding labels, based on the division of the chip feature region group. The purpose of this region marking is to establish a standard reference image (i.e., a baseline chip package image) with known feature location information, so that subsequent feature matching and positional deviation calculation can be performed with the image of the chip to be calibrated, providing a benchmark for package calibration.

[0098] S2. Use the pre-built visual positioning unit to acquire images of the chip to be calibrated to obtain the packaged image of the chip to be calibrated.

[0099] As is clear, the visual positioning unit refers to a device capable of capturing images of the chip to be calibrated, such as a high-resolution CCD industrial camera. The packaged image of the chip to be calibrated refers to a top-view image of the packaged chip to be calibrated and its packaging base, captured by the visual positioning unit, wherein the packaging base is of the same model as the aforementioned standard packaging base.

[0100] S3. Identify the chip feature regions in the chip package image to be calibrated to obtain a group of effective feature regions.

[0101] Understandably, the effective feature region group refers to a set of multiple effective feature regions, wherein the effective feature region refers to the image region corresponding to a certain chip feature region in the chip package image to be calibrated and the chip feature region group.

[0102] Specifically, the step of identifying chip feature regions from the chip package image to be calibrated to obtain a group of effective feature regions includes:

[0103] The image of the chip package to be calibrated is converted to grayscale to obtain a grayscale image of the chip to be calibrated. Then, grayscale threshold segmentation is performed on the grayscale image of the chip to be calibrated to obtain a segmented image of the chip to be calibrated, wherein the segmented image of the chip to be calibrated includes multiple pixels to be calibrated.

[0104] Chip contour extraction is performed on the segmented image of the chip to be calibrated to obtain the chip contour region;

[0105] The current center pixel group is identified based on the chip feature region group and the chip contour region. The current center pixel group includes multiple current center pixels, and each current center pixel corresponds one-to-one with a chip feature region.

[0106] Based on the current group of center pixels, perform region growing to obtain a group of effective feature regions.

[0107] It should be explained that the grayscale image of the chip to be calibrated refers to the image of the chip package after grayscale conversion. The segmented image of the chip to be calibrated refers to the binary image obtained after performing grayscale thresholding on the grayscale image of the chip to be calibrated. Grayscale thresholding refers to an image processing technique that divides the pixels in the chip package image into different regions (such as foreground and background). Optionally, algorithms such as the Otsu method or adaptive thresholding method are used for grayscale thresholding. The pixels to be calibrated refer to the pixels in the segmented image of the chip to be calibrated. The chip contour region refers to the rectangular region enclosed by the overall outer contour of the chip to be calibrated in the segmented image of the chip to be calibrated. Chip contour extraction refers to the process of detecting and connecting the edges of the target object (i.e., the chip to be calibrated) from the segmented image of the chip to be calibrated. Optionally, algorithms such as Canny edge detection and Sobel operator combined with contour search algorithm are used for chip contour extraction. The current center pixel group refers to a set of multiple current center pixels. The current center pixel refers to the pixel in the segmented image of the chip to be calibrated that corresponds to the geometric center of a certain chip feature region.

[0108] Specifically, the identification of the current center pixel group based on the chip feature region group and the chip contour region includes:

[0109] The standard contour region is extracted based on the reference chip packaging image, and a coordinate system is constructed for the standard contour region to obtain the standard contour coordinate system.

[0110] The chip contour region is scaled and transformed based on the standard contour region to obtain the target contour region, wherein the coordinate system of the target contour region is the same as the coordinate system of the standard contour region.

[0111] Extract chip feature regions sequentially from the chip feature region group and identify the standard region center of the extracted chip feature regions;

[0112] Determine the standard center coordinates of the standard region center based on the standard contour coordinate system;

[0113] Determine the current center pixel point in the target contour region that corresponds to the standard center coordinates;

[0114] The current center pixel points corresponding to each chip feature region are summarized to obtain the current center pixel point group.

[0115] It should be explained that the standard contour region refers to the rectangular area enclosed by the outer contour of the same model chip in the reference chip package image. The standard contour coordinate system refers to a two-dimensional coordinate system constructed on the standard contour region. Since the chip to be calibrated has the same size and structure as the chip of the same model, the purpose of constructing this standard contour coordinate system is to enable accurate coordinate mapping and positional correspondence between the target contour region (from the chip to be calibrated) and the standard contour region. Therefore, this solution does not limit the specific construction position of the standard contour coordinate system. The target contour region refers to the chip contour region after scaling transformation. Scaling transformation refers to performing an affine transformation (such as translation, rotation, or scaling) on ​​the chip contour region to make it the same as the standard contour region in size, orientation, and position. The coordinate system in the target contour region is constructed at the same position as the standard contour coordinate system in the standard contour region. The center of the standard region refers to the geometric center of the chip feature region. The standard center coordinates refer to the position coordinates of the center of the standard region in the standard contour coordinate system. The above-mentioned determination of the current center pixel point corresponding to the standard center coordinates in the target contour region means: recording the pixel point in the target contour region with the same coordinates as the standard center coordinates as the current center pixel point.

[0116] Specifically, the step of performing region growing based on the current group of center pixels to obtain a group of effective feature regions includes:

[0117] Extract the current center pixel from the current center pixel group in sequence, and record the extracted current center pixel as the starting pixel;

[0118] Get the starting pixel value of the starting pixel;

[0119] In the segmented image of the chip to be calibrated, identify the set of neighboring pixels corresponding to the starting pixel, where the set of neighboring pixels includes multiple neighboring pixels.

[0120] Obtain the set of neighboring pixel values ​​of the set of neighboring pixels, where the neighboring pixel values ​​in the set of neighboring pixel values ​​correspond one-to-one with the neighboring pixels;

[0121] Based on the adjacent pixel value set and the starting pixel value, the same pixel value is judged to obtain the pixel judgment result, where the pixel judgment result is either that there are the same pixel values ​​or that there are no the same pixel values;

[0122] If the pixel discrimination result is that there are identical pixel values, then the pixel group in the adjacent pixel set is identified based on the adjacent pixel value set and the starting pixel value.

[0123] The pixel group within the domain is added to the pre-constructed original feature region pixel set to obtain the target feature region pixel set;

[0124] The target feature region pixel set and the domain pixel group are respectively used as the original feature region pixel set and the current center pixel group, and the step of extracting the current center pixel in the current center pixel group is returned until the pixel discrimination result is that there are no identical pixel values.

[0125] If the pixel discrimination result is that there are no identical pixel values, then the original feature region pixel set is recorded as the effective feature region pixel set;

[0126] The effective feature region is divided according to the set of pixels in the effective feature region;

[0127] The effective feature regions corresponding to each current center pixel are summarized to obtain the effective feature region group.

[0128] It should be explained that the starting pixel value refers to the pixel value corresponding to the starting pixel. The neighboring pixel set refers to a set of multiple neighboring pixels, where a neighboring pixel is a pixel in the segmented image of the chip to be calibrated that is directly adjacent to the starting pixel (a four-connected or eight-connected neighborhood), for example: pixels in the four directions above, below, left, and right of the starting pixel (four-neighborhood), or pixels in eight directions including the four diagonal directions (eight-neighborhood). The neighboring pixel value set refers to the set of pixel values ​​(i.e., neighboring pixel values) of each neighboring pixel in the neighboring pixel set. The pixel discrimination result refers to the result obtained after discrimination based on the same pixel value, where "having the same pixel value" means that there is a neighboring pixel value with the same value as the starting pixel value in the neighboring pixel value set, and "not having the same pixel value" means that there is no neighboring pixel value with the same value as the starting pixel value in the neighboring pixel value set. If the pixel identification result indicates the existence of identical pixel values, it means that the effective feature region can be expanded from the starting pixel point to these adjacent pixels with the same pixel values. If the pixel identification result indicates the absence of identical pixel values, it means that with the current starting pixel point as the growth point, it is impossible to find any adjacent pixels with the same pixel values ​​connected to it, and the growth process of the corresponding effective feature region ends.

[0129] Furthermore, the "domain-internal pixel group" refers to a set of multiple domain-internal pixels. A domain-internal pixel refers to an adjacent pixel whose value is the same as the starting pixel value. The aforementioned identification of the domain-internal pixel group based on the adjacent pixel value set and the starting pixel value means: identifying adjacent pixel values ​​with the same value as the starting pixel value in the adjacent pixel value set, and recording the adjacent pixel corresponding to this adjacent pixel value as a domain-internal pixel. All domain-internal pixels then form the domain-internal pixel group. The "original feature region pixel set" refers to the target feature region pixel set before the previous return step. The "return step" refers to the subsequent return to the step of sequentially extracting the current center pixel from the current center pixel group. If the subsequent return step has not yet been executed, the original feature region pixel set only contains the starting pixel. The "target feature region pixel set" refers to the new pixel set obtained after supplementing the domain-internal pixel group to the original feature region pixel set. This target feature region pixel set represents the set of all pixels belonging to the same valid feature region that have been aggregated from the starting center pixel. The above-mentioned division of effective feature regions based on the set of effective feature region pixels refers to: in the segmented image of the chip to be calibrated, marking and extracting the connected image sub-regions occupied by all pixels belonging to the same set of effective feature region pixels, as an independent effective feature region.

[0130] S4. Extract effective feature regions sequentially from the effective feature region group, and identify the matching feature regions corresponding to the extracted effective feature regions in the chip feature region group of the reference chip package image.

[0131] It is clear that the feature region to be matched refers to the chip feature region in the reference chip package image that represents the same feature region as the extracted effective feature region.

[0132] S5. Perform encapsulation deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector.

[0133] Understandably, the target deviation correction vector refers to a mathematical vector used to correct the position and angle deviations of the chip package to be calibrated.

[0134] In detail, the step of performing encapsulation deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector includes:

[0135] Based on the extracted effective feature regions, the packaging region is selected in the grayscale image of the chip to be calibrated to obtain the effective packaging region. Based on the feature regions to be matched, the reference packaging region is selected in the reference chip packaging image.

[0136] Pyramid downsampling is performed on the effective encapsulation region to obtain an effective pyramid image set, which includes multiple effective pyramid images;

[0137] Obtain a reference pyramid image set for the reference encapsulation region, wherein the reference pyramid image set includes multiple reference pyramid images, and each reference pyramid image corresponds one-to-one with a valid pyramid image;

[0138] The effective pyramid images are extracted sequentially from the effective pyramid image set, and the extracted effective pyramid images are recorded as the top effective pyramid images. The top effective pyramid images are then transformed using a preset initial deviation correction vector to obtain the transformed top pyramid images.

[0139] Identify the reference pyramid top-level image corresponding to the valid pyramid top-level image in the reference pyramid image set;

[0140] The target deviation correction vector is constructed based on the top-level image of the baseline pyramid and the top-level image of the transformed pyramid.

[0141] It should be explained that the effective packaging region refers to a local image region centered on the effective feature region within the grayscale image of the chip to be calibrated. This effective packaging region is selected by defining a square area with a fixed radius and the geometric center of the effective feature region as the origin. The effective feature region is only a portion of the chip to be calibrated, and the effective packaging region includes both the packaging base and the effective feature region. The reference packaging region refers to a local image region centered on the feature region to be matched within the packaged image of a reference chip. The selection method for the reference packaging region is the same as that for the effective packaging region. The effective pyramid image set refers to a collection of multiple effective pyramid images. Each effective pyramid image refers to an image at different resolutions obtained after pyramid downsampling the effective packaging region. Different levels (i.e., different positions within the effective pyramid image set) represent approximations of the effective packaging region at different scales (i.e., different resolutions). The pyramid downsampling described above can be performed using a Gaussian pyramid algorithm or method. The reference pyramid image set refers to a collection of multiple reference pyramid images, where each reference pyramid image refers to an image at a different resolution obtained by pyramid downsampling a reference encapsulation region. The reference pyramid image set is acquired in the same way as the valid pyramid image set.

[0142] Furthermore, the initial deviation correction vector refers to the encapsulated deviation correction vector obtained after deviation correction of the next layer of effective pyramid image. The numerical categories contained in the initial deviation correction vector are the same as those contained in the subsequent region correction data, that is, both contain: region correction horizontal axis displacement, region correction vertical axis displacement, and region clockwise rotation. The next layer of effective pyramid image refers to the effective pyramid image that is arranged one position after the current effective pyramid image in the effective pyramid image set. If the effective pyramid image extracted at this time is the effective pyramid image that is arranged at the first position in the effective pyramid image set, then the initial deviation correction vector is a unit vector. That is, when the image transformation is performed on the top layer of the effective pyramid using the initial deviation correction vector, the top layer of the effective pyramid will not change. The transformed pyramid top-level image refers to the image obtained after transforming the effective pyramid top-level image. The image transformation using a preset initial deviation correction vector involves translating and rotating the region belonging to the chip to be detected (i.e., the effective feature region in the effective pyramid top-level image) according to the region correction horizontal axis displacement, region correction vertical axis displacement, and region clockwise rotation amount in the initial deviation correction vector. The effective pyramid top-level image after translation and rotation is the transformed pyramid top-level image. The reference pyramid top-level image refers to the reference pyramid top-level image that has the same arrangement position as the effective pyramid top-level image in the reference pyramid top-level image set.

[0143] Specifically, the step of constructing the target deviation correction vector based on the reference pyramid top-level image and the transformed pyramid top-level image includes:

[0144] The top image of the transform pyramid is encapsulated with the top image of the base pyramid to identify the deviation, and the deviation identification result is obtained, which indicates whether the deviation exists or not.

[0145] If the deviation identification result has a deviation, then obtain the encapsulated deviation correction vector;

[0146] The encapsulated deviation correction vector is used as the initial deviation correction vector, and the step of sequentially extracting valid pyramid images in the valid pyramid image set is returned until the deviation identification result is that there is no deviation.

[0147] If the deviation identification result is that there is no deviation, then the initial deviation correction vector is recorded as the target deviation correction vector.

[0148] It should be explained that the deviation identification result refers to the result obtained after encapsulation deviation identification. Here, "existence of deviation" means there is a positional deviation between the reference pyramid top-level image and the transformed pyramid top-level image; "no deviation" means there is no positional deviation between the reference pyramid top-level image and the transformed pyramid top-level image. When the deviation identification result shows a deviation, it indicates that the effective feature region has shifted positionally at the current resolution. In this case, to more accurately obtain the specific value of the deviation, deviation identification of the next layer of effective pyramid images is required. When the deviation identification result shows no deviation, it indicates that the effective feature region has not shifted positionally at the current resolution, and deviation identification of the next layer of effective pyramid images is not performed. The encapsulation deviation correction vector refers to the numerical vector composed of the cumulative values ​​of the horizontal axis displacement, vertical axis displacement, and clockwise rotation of the region correction data obtained when the deviation identification result has a deviation. For example, when the deviation identification result has a deviation, region correction data A (XA, YA, SA), region correction data B (XB, YB, SB), and region correction data C (XC, YC, SC) have been obtained. Among them, XA, YA, SA, XB, YB, SB, XC, YC, and SC represent the horizontal axis displacement, vertical axis displacement, and clockwise rotation of the region correction data A, B, and C, respectively. Then, these values ​​are accumulated to obtain the encapsulation deviation correction vector as (XA+XB+XC, YA+YB+YC, SA+SB+SC).

[0149] Furthermore, this scheme obtains effective pyramid images at different resolutions through the pyramid downsampling operation. Then, following the order of resolution from coarse to fine (i.e., the effective pyramid images are arranged from front to back), the effective pyramid images are sequentially subjected to deviation discrimination, and the correction data for the region requiring deviation adjustment in each effective pyramid image is calculated. Deviation discrimination is stopped in time when no deviation occurs, and the accumulated encapsulated deviation correction vector is calculated. Compared with the traditional scheme of performing global search and matching directly on the original high-resolution image, the pyramid layering strategy adopted in this scheme can perform preliminary matching and deviation correction on the low-resolution (coarse) effective pyramid images. It has less computation and is faster, and can quickly eliminate large offsets. Then, the preliminary correction result (i.e., the encapsulated deviation correction vector) is used as the initial deviation correction vector for matching the next layer of higher-resolution images, thereby performing more refined adjustments. This iterative process from coarse to fine avoids large-scale and time-consuming exhaustive searches at the fine scale. While ensuring the final calibration accuracy, it reduces the overall computation time. Secondly, the early termination mechanism (stopping when there is no deviation in a certain layer) further optimizes computational resources and avoids unnecessary fine calculations.

[0150] In detail, the step of using the top-level image of the reference pyramid to perform encapsulation deviation identification on the top-level image of the transformed pyramid to obtain the deviation identification result includes:

[0151] The set of transformed pixels in the top layer of the transform pyramid image is identified, wherein the set of transformed pixels includes multiple transformed pixels;

[0152] Transformed pixels are extracted sequentially from the set of transformed pixels, and the neighborhood of transformed pixels is identified in the top layer image of the transformation pyramid based on the extracted transformed pixels and the preset matching template size.

[0153] Region matching is performed on the neighborhood of transformed pixels based on the top-level image of the baseline pyramid to obtain the region matching degree.

[0154] Summarize the region matching degree corresponding to each transformed pixel to obtain the region matching degree set;

[0155] Identify the maximum regional deviation value based on the regional matching degree set, where the maximum regional deviation value is the minimum value in the regional matching degree set;

[0156] Obtain the target transformation region and target reference region corresponding to the maximum regional deviation value;

[0157] The target transformation region is solved by using the target reference region to obtain the region correction data, which includes: the region correction horizontal axis displacement, the region correction vertical axis displacement, and the region clockwise rotation.

[0158] The regional correction amount is calculated based on the regional correction data. If the regional correction amount is greater than the preset standard correction amount, the existence of deviation is recorded as the deviation identification result; otherwise, the absence of deviation is recorded as the deviation identification result.

[0159] It should be explained that the "transformed pixel set" refers to a collection of multiple transformed pixels, where each transformed pixel refers to a pixel in the top-level image of the transformation pyramid. The "matching template size" refers to the width and height of the local image block used for sliding matching on the top-level image of the transformation pyramid, for example: 16 pixels × 16 pixels, 32 pixels × 32 pixels. The "transformed pixel neighborhood" refers to the local image region extracted from the top-level image of the transformation pyramid, centered on the currently extracted transformed pixel and within the range of the matching template size. The "region matching degree" refers to the similarity between the transformed pixel neighborhood and the most similar part of the reference top-level image of the pyramid; a higher region matching degree indicates a greater similarity. The "maximum region deviation value" refers to the region matching degree with the smallest value in the region matching degree set; this maximum region deviation value represents the maximum deviation between the top-level image of the transformation pyramid and the reference top-level image of the pyramid. The "target transformed region" and "target reference region" refer to the transformed pixel neighborhood and the traversed reference image region corresponding to the maximum region deviation value, respectively. The region correction data refers to the set of region correction horizontal axis displacement, region correction vertical axis displacement, and region clockwise rotation. The region correction horizontal axis displacement refers to the horizontal axis translation required to transform the position of the target transformation region to the position of the target reference region. The region correction vertical axis displacement refers to the vertical axis translation required to transform the position of the target transformation region to the position of the target reference region. The region clockwise rotation refers to the clockwise rotation angle required to transform the position of the target transformation region to the position of the target reference region. All of the above position transformations are based on the coordinate system of the chip package image to be calibrated.

[0160] Furthermore, the aforementioned method of solving the target transformation region using the target reference region is as follows: The spatial transformation relationship between the target transformation region and the target reference region is calculated using an image registration algorithm (such as feature-based registration or region-based registration). A further method involves extracting conjugate feature point pairs between the target transformation region and the target reference region to solve for an optimal transformation matrix (such as an affine transformation or similarity transformation matrix). The translation and rotation parameters contained in this transformation matrix constitute the region correction data. The aforementioned image registration algorithm is existing technology and will not be elaborated further here. The region correction amount refers to the degree of correction required when correcting according to the region correction data. The smaller the region correction amount, the smaller the degree of correction required. When the degree of correction is less than a certain threshold (i.e., the standard correction amount), to avoid unnecessary adjustments to small, negligible noise and to improve the efficiency of the overall calibration process, no correction is required; the absence of deviation is recorded as a deviation identification result. The aforementioned standard correction amount refers to a preset threshold used to determine whether a deviation needs correction. This standard correction amount is set by the user according to the required calibration precision. The calculation method for the above-mentioned area correction amount is as follows:

[0161] ;

[0162] in, Indicates the regional correction amount. This represents a normalization function, such as minimum-maximum normalization. These represent the horizontal displacement of the region, the vertical displacement of the region, and the clockwise rotation of the region, respectively.

[0163] In detail, the process of performing region matching on the transformed pixel neighborhood based on the top-level image of the benchmark pyramid to obtain the region matching degree includes:

[0164] Feature extraction is performed on the neighborhood of the transformed pixels to obtain the features of the transformed region.

[0165] Based on the matching template size, the top layer image of the baseline pyramid is traversed to obtain the traversed baseline image region.

[0166] Obtain the traversal region features of the reference image region, and calculate the region matching degree based on the transformed region features and the traversal region features.

[0167] Return to the step of traversing the top-level image of the baseline pyramid based on the matching template size, until the traversal is complete;

[0168] Summarize the regional matching scores to obtain the regional matching score set, identify the maximum matching score in the regional matching score set, and record the maximum matching score as the regional matching score.

[0169] It should be explained that the transformed region features refer to the grayscale features of the transformed pixel neighborhood, such as pixel grayscale value matrices, grayscale histograms, or directional gradient histograms, etc., which can be extracted using algorithms such as SIFT, SURF, or ORB. The traversal of the reference image region refers to sequentially extracting a local image region from the top-level image of the reference pyramid using a sliding window of the same size as the matching template. Traversing the top-level image of the reference pyramid based on the matching template size means: using the matching template size as a window, starting from the top left corner of the top-level image of the reference pyramid, sliding the window sequentially to the right and down according to a set step size (e.g., 1 pixel), and extracting an image block of the same size as the window at each new position; this image block constitutes a traversed reference image region. The traversed region features refer to the grayscale features of the traversed reference image region, and the extraction method for these traversed region features is the same as the extraction method for the transformed region features. The region matching degree refers to a numerical value that quantifies the similarity between the transformed region features and the traversed region features. The region matching degree is calculated as follows: the transformed region features and the traversed region features are each converted into numerical vectors with the same value. Then, the cosine value of the vector between the numerical vectors corresponding to the transformed region features and the numerical vectors corresponding to the traversed region features is calculated, and this cosine value is used as the region matching degree. "Completed traversal" means that the sliding window has traversed all possible positions in the top layer image of the baseline pyramid, i.e., it is no longer possible to capture a new, complete traversed baseline image region with the same size as the matching template.

[0170] S6. Summarize the target deviation correction vectors corresponding to each effective feature region to obtain the target deviation correction vector group. Perform vector meanization on the target deviation correction vector group to obtain the average deviation correction vector. Generate a chip packaging calibration report based on the average deviation correction vector to complete the electronic chip packaging calibration based on vision positioning.

[0171] As is clear, the average deviation correction vector refers to the vector obtained by averaging the target deviation correction vector group. Vector averaging involves taking the average of the vector elements at the same position in each target deviation correction vector group; the average of these vector elements at multiple different positions constitutes the average deviation correction vector. The chip packaging calibration report refers to a data file composed of the average deviation correction vector. Operators can use this report to adjust the parameters of packaging equipment (such as pick-and-place machines and die bonders) to correct the pose deviations of current or subsequent chip packages of the same type, thereby improving production accuracy and yield.

[0172] To address the problems described in the background section, this invention first obtains a reference chip package image based on the chip to be calibrated. This step involves precisely packaging a chip of the same model with a standard packaging base and capturing the image, constructing a physically accurate reference image containing known feature location information, rather than relying on theoretical drawings or CAD models. This provides a high-fidelity reference standard for subsequent visual comparison, avoiding systematic errors introduced by differences between models and physical objects. Furthermore, this solution performs chip feature region identification on the chip package image to be calibrated, obtaining a group of effective feature regions. This step, through image processing, contour extraction, coordinate system mapping, and a region growing algorithm, can automatically and accurately segment and identify multiple feature regions corresponding to the reference image from the chip image to be calibrated. This method ensures the reliability of feature recognition, providing a foundation for subsequent precise... The accurate alignment lays the foundation. Finally, based on the feature regions to be matched and the extracted effective feature regions, packaging deviation calibration is performed to obtain the target deviation correction vector. This step adopts a coarse-to-fine pyramid hierarchical matching and iterative correction strategy. Compared with the traditional approach of directly performing global search matching on the original high-resolution image, this method first performs fast matching and preliminary deviation correction at the low-resolution image layer, which has a small computational load and can quickly eliminate large offsets. Then, the preliminary result is used as the initial value for iterative fine adjustment at the higher-resolution image layer. This strategy avoids large-scale and time-consuming exhaustive searches at a fine scale, reducing the overall computation time while ensuring the final calibration accuracy. At the same time, the introduced early termination mechanism (stopping when there is no deviation in a certain layer) further optimizes computational resources and avoids unnecessary fine calculations. Therefore, this invention can improve the computational efficiency and overall stability of electronic chip packaging calibration and reduce sensitivity to irrelevant image noise.

[0173] like Figure 2 The diagram shown is a functional block diagram of an electronic chip packaging and calibration system based on visual positioning provided in an embodiment of the present invention.

[0174] The vision-based electronic chip packaging calibration system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the vision-based electronic chip packaging calibration system 100 may include a packaging instruction receiving module 101, a feature region recognition module 102, a standard region matching module 103, and a calibration report generation module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0175] The packaging instruction receiving module 101 is used to receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0176] The feature region recognition module 102 is used to acquire images of the chip to be calibrated using a pre-built visual positioning unit to obtain a packaged image of the chip to be calibrated, and to identify chip feature regions in the packaged image of the chip to be calibrated to obtain a group of effective feature regions.

[0177] The standard region matching module 103 is used to extract effective feature regions sequentially from the effective feature region group and identify the feature region to be matched corresponding to the extracted effective feature region in the chip feature region group in the reference chip package image.

[0178] The calibration report generation module 104 is used to perform packaging deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector, summarize the target deviation correction vector corresponding to each effective feature region to obtain the target deviation correction vector group, and generate a chip packaging calibration report based on the average deviation correction vector.

[0179] In detail, the modules in the vision-based electronic chip packaging calibration system 100 described in this embodiment of the invention employ the same methods as described above during use. Figure 1 The method used is the same as the vision-based positioning method for electronic chip packaging calibration described above, and it can produce the same technical effect, so it will not be repeated here.

[0180] like Figure 3 The diagram shown is a structural schematic of an electronic device that implements a vision-based positioning-based electronic chip packaging calibration method according to an embodiment of the present invention.

[0181] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a visual positioning-based electronic chip packaging calibration method program.

[0182] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a vision-based electronic chip packaging calibration method program, but also to temporarily store data that has been output or will be output.

[0183] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a visual positioning-based electronic chip packaging calibration method program) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0184] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0185] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0186] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0187] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0188] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0189] The vision-based positioning electronic chip packaging calibration method program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0190] Receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0191] The pre-built visual positioning unit is used to acquire images of the chip to be calibrated, thereby obtaining the packaged image of the chip to be calibrated;

[0192] The chip feature region is identified from the chip package image to be calibrated to obtain a group of effective feature regions;

[0193] Effective feature regions are extracted sequentially from the effective feature region group, and the matching feature regions corresponding to the extracted effective feature regions are identified from the chip feature region group in the reference chip package image.

[0194] Encapsulation deviation calibration is performed based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector;

[0195] The target deviation correction vectors corresponding to each effective feature region are summarized to obtain the target deviation correction vector group. The target deviation correction vector group is then averaged to obtain the average deviation correction vector. Based on the average deviation correction vector, a chip packaging calibration report is generated, thus completing the vision-based positioning-based electronic chip packaging calibration.

[0196] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0197] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0198] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0199] Receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group.

[0200] The pre-built visual positioning unit is used to acquire images of the chip to be calibrated, thereby obtaining the packaged image of the chip to be calibrated;

[0201] The chip feature region is identified from the chip package image to be calibrated to obtain a group of effective feature regions;

[0202] Effective feature regions are extracted sequentially from the effective feature region group, and the matching feature regions corresponding to the extracted effective feature regions are identified from the chip feature region group in the reference chip package image.

[0203] Encapsulation deviation calibration is performed based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector;

[0204] The target deviation correction vectors corresponding to each effective feature region are summarized to obtain the target deviation correction vector group. The target deviation correction vector group is then averaged to obtain the average deviation correction vector. Based on the average deviation correction vector, a chip packaging calibration report is generated, thus completing the vision-based positioning-based electronic chip packaging calibration.

[0205] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0206] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0207] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0208] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0209] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for calibrating electronic chip packaging based on visual positioning, characterized in that, The method includes: Receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group. The pre-built visual positioning unit is used to acquire images of the chip to be calibrated, thereby obtaining the packaged image of the chip to be calibrated; The chip feature region is identified from the chip package image to be calibrated to obtain a group of effective feature regions; Effective feature regions are extracted sequentially from the effective feature region group, and the matching feature regions corresponding to the extracted effective feature regions are identified from the chip feature region group in the reference chip package image. Encapsulation deviation calibration is performed based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector; The target deviation correction vectors corresponding to each effective feature region are summarized to obtain the target deviation correction vector group. The target deviation correction vector group is then averaged to obtain the average deviation correction vector. Based on the average deviation correction vector, a chip packaging calibration report is generated, thus completing the vision-based positioning-based electronic chip packaging calibration.

2. The electronic chip packaging calibration method based on vision positioning as described in claim 1, characterized in that, The step of obtaining the reference chip package image based on the chip to be calibrated includes: Obtain chips of the same model based on the chip to be calibrated; A standard chip packaging base is obtained by precisely packaging chips of the same model using a preset standard packaging base. The standard chip packaging base contains chips of the same model. The chip is photographed using a visual positioning unit on a standard chip packaging base to obtain an initial chip packaging image; Chips of the same model are divided into feature regions to obtain chip feature region groups, where each chip feature region group includes multiple chip feature regions. The initial chip package image is marked with regions based on the chip feature region group to obtain the reference chip package image.

3. The method of claim 2, wherein the method further comprises: The process of identifying chip feature regions from the chip package image to be calibrated, resulting in a group of effective feature regions, includes: The image of the chip package to be calibrated is converted to grayscale to obtain a grayscale image of the chip to be calibrated. Then, grayscale threshold segmentation is performed on the grayscale image of the chip to be calibrated to obtain a segmented image of the chip to be calibrated, wherein the segmented image of the chip to be calibrated includes multiple pixels to be calibrated. Chip contour extraction is performed on the segmented image of the chip to be calibrated to obtain the chip contour region; The current center pixel group is identified based on the chip feature region group and the chip contour region. The current center pixel group includes multiple current center pixels, and each current center pixel corresponds one-to-one with a chip feature region. Based on the current group of center pixels, perform region growing to obtain a group of effective feature regions.

4. The electronic chip packaging calibration method based on vision positioning as described in claim 3, characterized in that, The method of identifying the current center pixel group based on chip feature region groups and chip contour regions includes: The standard contour region is extracted based on the reference chip packaging image, and a coordinate system is constructed for the standard contour region to obtain the standard contour coordinate system. The chip contour region is scaled and transformed based on the standard contour region to obtain the target contour region, wherein the coordinate system of the target contour region is the same as the coordinate system of the standard contour region. Extract chip feature regions sequentially from the chip feature region group and identify the standard region center of the extracted chip feature regions; Determine the standard center coordinates of the standard region center based on the standard contour coordinate system; Determine the current center pixel point in the target contour region that corresponds to the standard center coordinates; The current center pixel points corresponding to each chip feature region are summarized to obtain the current center pixel point group.

5. The method of claim 4, wherein the method further comprises: The step of performing region growing based on the current group of center pixels to obtain a group of effective feature regions includes: Extract the current center pixel from the current center pixel group in sequence, and record the extracted current center pixel as the starting pixel; Get the starting pixel value of the starting pixel; In the segmented image of the chip to be calibrated, identify the set of neighboring pixels corresponding to the starting pixel, where the set of neighboring pixels includes multiple neighboring pixels. Obtain the set of neighboring pixel values ​​of the set of neighboring pixels, where the neighboring pixel values ​​in the set of neighboring pixel values ​​correspond one-to-one with the neighboring pixels; Based on the adjacent pixel value set and the starting pixel value, the same pixel value is judged to obtain the pixel judgment result, where the pixel judgment result is either that there are the same pixel values ​​or that there are no the same pixel values; If the pixel discrimination result is that there are identical pixel values, then the pixel group in the adjacent pixel set is identified based on the adjacent pixel value set and the starting pixel value. The pixel group within the domain is added to the pre-constructed original feature region pixel set to obtain the target feature region pixel set; The target feature region pixel set and the domain pixel group are respectively used as the original feature region pixel set and the current center pixel group, and the step of extracting the current center pixel in the current center pixel group is returned until the pixel discrimination result is that there are no identical pixel values. If the pixel discrimination result is that there are no identical pixel values, then the original feature region pixel set is recorded as the effective feature region pixel set; The effective feature region is divided according to the set of pixels in the effective feature region; The effective feature regions corresponding to each current center pixel are summarized to obtain the effective feature region group.

6. The vision positioning based electronic chip package calibration method of claim 5, wherein, The step of performing encapsulation deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector includes: Based on the extracted effective feature regions, the packaging region is selected in the grayscale image of the chip to be calibrated to obtain the effective packaging region. Based on the feature regions to be matched, the reference packaging region is selected in the reference chip packaging image. Pyramid downsampling is performed on the effective encapsulation region to obtain an effective pyramid image set, which includes multiple effective pyramid images; Obtain a reference pyramid image set for the reference encapsulation region, wherein the reference pyramid image set includes multiple reference pyramid images, and each reference pyramid image corresponds one-to-one with a valid pyramid image; The effective pyramid images are extracted sequentially from the effective pyramid image set, and the extracted effective pyramid images are recorded as the top effective pyramid images. The top effective pyramid images are then transformed using a preset initial deviation correction vector to obtain the transformed top pyramid images. Identify the reference pyramid top-level image corresponding to the valid pyramid top-level image in the reference pyramid image set; The target deviation correction vector is constructed based on the top-level image of the baseline pyramid and the top-level image of the transformed pyramid.

7. The vision positioning based electronic chip package calibration method of claim 6, wherein, The construction of the target deviation correction vector based on the top-level image of the baseline pyramid and the top-level image of the transformed pyramid includes: The top image of the transform pyramid is encapsulated with the top image of the base pyramid to identify the deviation, and the deviation identification result is obtained, which indicates whether the deviation exists or not. If the deviation identification result has a deviation, then obtain the encapsulated deviation correction vector; The encapsulated deviation correction vector is used as the initial deviation correction vector, and the step of sequentially extracting valid pyramid images in the valid pyramid image set is returned until the deviation identification result is that there is no deviation. If the deviation identification result is that there is no deviation, then the initial deviation correction vector is recorded as the target deviation correction vector.

8. The vision positioning based electronic chip package calibration method of claim 7, wherein, The method of using the top-level image of the baseline pyramid to perform encapsulation deviation identification on the top-level image of the transformed pyramid, and obtaining the deviation identification result, includes: The set of transformed pixels in the top layer of the transform pyramid image is identified, wherein the set of transformed pixels includes multiple transformed pixels; Transformed pixels are extracted sequentially from the set of transformed pixels, and the neighborhood of transformed pixels is identified in the top layer image of the transformation pyramid based on the extracted transformed pixels and the preset matching template size. Region matching is performed on the neighborhood of transformed pixels based on the top-level image of the baseline pyramid to obtain the region matching degree. Summarize the region matching degree corresponding to each transformed pixel to obtain the region matching degree set; Identify the maximum regional deviation value based on the regional matching degree set, where the maximum regional deviation value is the minimum value in the regional matching degree set; Obtain the target transformation region and target reference region corresponding to the maximum regional deviation value; The target transformation region is solved by using the target reference region to obtain the region correction data, which includes: the region correction horizontal axis displacement, the region correction vertical axis displacement, and the region clockwise rotation. The regional correction amount is calculated based on the regional correction data. If the regional correction amount is greater than the preset standard correction amount, the existence of deviation is recorded as the deviation identification result; otherwise, the absence of deviation is recorded as the deviation identification result.

9. The vision positioning based electronic chip package calibration method of claim 8, wherein, The method of performing region matching on the transformed pixel neighborhood based on the top-level image of the benchmark pyramid to obtain the region matching degree includes: Feature extraction is performed on the neighborhood of the transformed pixels to obtain the features of the transformed region. Based on the matching template size, the top layer image of the baseline pyramid is traversed to obtain the traversed baseline image region. Obtain the traversal region features of the reference image region, and calculate the region matching degree based on the transformed region features and the traversal region features. Return to the step of traversing the top-level image of the baseline pyramid based on the matching template size, until the traversal is complete; Summarize the regional matching scores to obtain the regional matching score set, identify the maximum matching score in the regional matching score set, and record the maximum matching score as the regional matching score.

10. An electronic chip package calibration system based on vision positioning, characterized in that, The system includes: The packaging instruction receiving module is used to receive electronic chip packaging instructions, identify the chip to be calibrated based on the electronic chip packaging instructions, and obtain a reference chip packaging image based on the chip to be calibrated, wherein the reference chip packaging image includes a chip feature region group. The feature region recognition module is used to acquire images of the chip to be calibrated using a pre-built visual positioning unit, obtain a packaged image of the chip to be calibrated, and perform chip feature region recognition on the packaged image of the chip to be calibrated to obtain a group of effective feature regions. The standard region matching module is used to extract effective feature regions sequentially from the effective feature region group and identify the feature region to be matched corresponding to the extracted effective feature region in the chip feature region group in the reference chip package image. The calibration report generation module is used to perform packaging deviation calibration based on the feature region to be matched and the extracted effective feature region to obtain the target deviation correction vector. It summarizes the target deviation correction vectors corresponding to each effective feature region to obtain the target deviation correction vector group, and generates a chip packaging calibration report based on the average deviation correction vector.

Images