A detection method based on complex chip type IPM

Abstract

The application discloses a kind of detection methods based on complex chip type IPM, include the following steps: S1: bright-dark image acquisition step;S2: chip positioning step;S3: wire detection step;S4: chip surface detection step;S5: position detection step;S6: comprehensive determination step.The application has the following beneficial effects: 1, the IPM detection method is a set of general detection process, combined with multiple sub-algorithms, adapt to IPM product characteristics, strong universality.2, solve the problem that all chips and soldering wires cannot be clearly imaged due to large height difference on product surface.3, detection process is flexible, and the detection process can be established according to different types of chips.4, can adapt to different number of chip detection, different types of soldering wire detection, etc., even more types are added, and the detection process and algorithm can be detected.5, the accurate position of gold wire and aluminum wire can be obtained before chip detection to shield, improve the accuracy of chip surface detection.

Description

Technical Field

[0001] This invention relates to the field of semiconductor packaging and testing technology, and in particular to a detection method based on IPM for complex chip types. Background Technology

[0002] With the continuous development of semiconductor technology, intelligent power modules (IPMs) have been widely used in industrial control, electric vehicles, and new energy power generation. IPM modules typically integrate multiple chip units, including power devices (such as IGBTs and MOSFETs), drive circuits, protection circuits, and control ICs, featuring compact structure, high functional integration, and high operating voltage and current ratings. Due to the complex internal structure and high thermal load of the modules, the manufacturing process of IPMs places higher demands on quality inspection.

[0003] Chip testing currently focuses on two main types: IPM and QFN. Testing items primarily include surface contamination and scratches, bent and broken solder wires, excessive or insufficient silver paste, and frame contamination and deformation. QFN products have a relatively simple composition, with most single modules consisting of a single chip and gold wires, making testing methods relatively mature. IPM products, however, have a complex internal structure, containing multiple chips of various types, as well as multiple gold and aluminum solder wires of different types. Testing accuracy reaches 10 micrometers, and product height differences are 1-2 mm, placing higher demands on equipment hardware and algorithms. The depth-of-field limitations of equipment lenses make it difficult to capture the entire chip in a single shot, and algorithms struggle to detect all defects in a single pass. This presents new challenges for IPM product testing methods. Summary of the Invention

[0004] According to an embodiment of the present invention, a detection method based on complex chip type IPM is provided, comprising the following steps:

[0005] Bright and dark image acquisition: By controlling the switchable combination light source and adjusting the height of the lens, clear images of all chips and solder lines on the product are acquired and the images are registered as templates.

[0006] Chip localization: Based on templates, the chip and frame in the current image are located using a template matching algorithm to obtain a transformation matrix representing the position and orientation;

[0007] Wire bonding inspection: Determine the inspection areas for aluminum and gold wires, extract the aluminum and gold wires within the inspection areas using image processing algorithms, and perform connectivity analysis to determine whether there are any breakage defects in the aluminum and gold wires;

[0008] Chip surface inspection: Receives aluminum and gold wire area information output from the wire bonding inspection step, performs shielding treatment on the areas of the chip surface covered by wires, obtains the real surface inspection area, extracts chip surface defects through threshold segmentation, and distinguishes dirt and scratches based on the shape characteristics of the defects.

[0009] Position detection: Based on the position information obtained in the chip positioning step, calculate the chip's position offset and rotation angle relative to the frame, and determine whether it is within the tolerance range;

[0010] Comprehensive judgment: Based on the results of the wire bonding inspection, chip surface inspection, and position inspection steps, if all inspection results are within the preset tolerance range, the product is judged to be qualified; otherwise, the product is judged to be unqualified.

[0011] Furthermore, the specific steps for obtaining the light and dark image are as follows:

[0012] To photograph the first type of chip, turn on the coaxial light in the combined light source and move the lens to a height where the first type of chip can be clearly photographed.

[0013] Adjust the type and brightness of the light source so that the chip in the image can be clearly imaged and the gray value of the chip surface is within the preset range;

[0014] Configure the currently adjusted chip image and register it as a template;

[0015] Repeat the above process until all chips that need to be photographed have completed image acquisition.

[0016] Furthermore, the specific steps for aluminum wire testing are as follows:

[0017] The detection area is selected by defining the detection location, and the coordinates of the detection frame are transformed by a transformation matrix to obtain the accurate coordinates of the aluminum wire detection area.

[0018] The aluminum wires and solder joints in the detection frame are identified and extracted using a threshold method. The extracted aluminum wire areas are then subjected to curvature analysis. If the curvature analysis meets the requirements, a connectivity analysis is performed to determine whether the aluminum wire is broken.

[0019] Furthermore, the extraction using the threshold method is as follows:

[0020] Based on the current grayscale value f(x,y) of the aluminum wire, select its corresponding grayscale threshold T;

[0021] Specify whether the aluminum wire is black or white. For black aluminum wire, specify the grayscale range as 0~T. For white aluminum wire, specify the grayscale range as T~255.

[0022] Furthermore, connectivity analysis specifically includes:

[0023] The extracted aluminum wire region is divided into multiple connected regions R1, R2, ..., Rn, satisfying...

[0024] ;

[0025] in, : Connect all connected regions R K The union of K=1 to n represents the foreground region of the entire image; R K The Kth connected region; : is the set of foreground pixels, that is, the set of all pixel coordinates with a value of 1; Any two distinct connected regions do not overlap;

[0026] If you find an area that runs through the first and second solder joints of the aluminum wire, it proves that the aluminum wire is not broken; otherwise, the aluminum wire is broken.

[0027] Furthermore, the specific steps for gold thread testing are as follows:

[0028] Gold wire solder joint inspection: Within the solder joint inspection area, the center and radius of the circular solder joint are located by extracting the edge of the circumferential caliper and fitting the circle using the least squares method.

[0029] Gold wire double solder joint detection: Within the double solder joint detection area, the irregular double solder joint positions are located using a grayscale template matching algorithm;

[0030] Gold wire extraction: Based on the location of the first and second solder joints, a gold wire detection area is generated. Within this area, the gold wire region is extracted using a threshold method, and connectivity analysis is performed to determine whether the gold wire is broken.

[0031] Furthermore, the specific inspection of a single solder joint on the gold wire is as follows:

[0032] Let the parameter of an initial circle be the center (x). c ,y c With radius r, N calipers are evenly placed on the circumference of the circle, centered at the circle's center. The coordinates of the center point of each caliper are:

[0033] ;

[0034] in, ; This is the initial angle;

[0035] On each caliper line segment, obtain the grayscale profile gi(s) and calculate its first derivative (gradient):

[0036] ;

[0037] This indicates the location where the grayscale change is most drastic, i.e., the edge point;

[0038] The coordinates of this edge point in the image are:

[0039] ;

[0040] All edge points Fit to an optimal circle:

[0041] ;

[0042] The final center and radius of the circle are obtained by minimizing the error function:

[0043] .

[0044] Furthermore, the specific inspection of the two solder joints of the gold wire is as follows:

[0045] Using a grayscale template matching algorithm, one or more standard two-weld-point templates are created, and weld points similar to the templates are searched in the two-weld-point detection area.

[0046] Gray-scale matching correlation coefficient:

[0047] ;

[0048] Where T: template grayscale set; The set of grayscale values ​​in the target image corresponding to the template position; Gray-scale dot product; : The grayscale energy product of the template and the target region;

[0049] Find the point with the maximum value as the matching result:

[0050] .

[0051] Furthermore, in the chip surface inspection step, when detecting chipped edges, it is necessary to shrink the chip surface according to the outline of the inspection, and extract and determine the defect threshold within this loop-shaped area.

[0052] Furthermore, the specific steps for location detection are as follows:

[0053] Based on the positioning information of the chip and the frame, settings are made, with the frame as the reference system and the chip as the object under test.

[0054] The offset range setting allows you to set the limit range by adjusting the chip's offset relative to the frame in the up, down, left, and right directions according to actual testing requirements.

[0055] The rotation range setting allows you to set the limit range by adjusting the rotation angle of the chip relative to the frame according to the actual testing requirements.

[0056] A detection method based on IPM of complex chip types according to an embodiment of the present invention has the following beneficial effects:

[0057] 1. This IPM detection method is a general detection process that combines multiple sub-algorithms, is adapted to the characteristics of IPM products, and has strong versatility.

[0058] 2. Solve the problem of unclear imaging of all chips and bonding wires due to large height differences on the product surface.

[0059] 3. The testing process is flexible and can be tailored to different types of chips.

[0060] 4. It can adapt to different numbers of chip inspections and different types of wire bonding inspections. Even if more types are added, the inspection process and algorithm can still perform the inspections.

[0061] 5. Before chip testing, the accurate positions of gold and aluminum wires can be obtained for shielding, improving the accuracy of chip surface testing.

[0062] It should be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further illustration of the claimed technology. Attached Figure Description

[0063] Figure 1 This is a schematic diagram of the structure of an IPM product in the existing technology.

[0064] Figure 2 This is a flowchart of a detection method based on complex chip type IPM according to an embodiment of the present invention.

[0065] Figure 3 This is a schematic diagram of the brightness and darkness image acquired during the detection process of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 1 .

[0066] Figure 4 This is a schematic diagram of the brightness and darkness image acquired during the detection process of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 2 .

[0067] Figure 5 This is a schematic diagram of the brightness and darkness image acquired during the detection process of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 3 .

[0068] Figure 6 This is a schematic diagram of the brightness and darkness image acquired during the detection process of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 4 .

[0069] Figure 7This is a schematic diagram of the brightness and darkness image acquired during the detection process of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 5 .

[0070] Figure 8 This is a schematic diagram of the detection results during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention.

[0071] Figure 9 This is a schematic diagram illustrating the shielding of the middle row of chips based on the location area of ​​the aluminum wire during the detection of a detection method for complex chip type IPM according to an embodiment of the present invention.

[0072] Figure 10 This is a schematic diagram illustrating the shielding of the chip based on the location area of ​​the gold wire during the detection of a detection method for complex chip type IPM according to an embodiment of the present invention.

[0073] Figure 11 This is a schematic diagram showing the bottom row of chips being shielded based on the location area of ​​the aluminum wire during a detection method based on a complex chip type IPM according to an embodiment of the present invention.

[0074] Figure 12 This is a schematic diagram showing the three chips on the right being shielded based on the location area of ​​the gold wire during a detection method based on a complex chip type IPM according to an embodiment of the present invention.

[0075] Figure 13 This is a schematic diagram of aluminum wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 1 .

[0076] Figure 14 This is a schematic diagram of aluminum wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 2 .

[0077] Figure 15 This is a schematic diagram of aluminum wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 3 .

[0078] Figure 16 This is a schematic diagram of aluminum wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 4 .

[0079] Figure 17 This is a schematic diagram of chip surface inspection during an inspection method based on complex chip type IPM according to an embodiment of the present invention. Figure 1 .

[0080] Figure 18 This is a schematic diagram of chip surface inspection during an inspection method based on complex chip type IPM according to an embodiment of the present invention. Figure 2 .

[0081] Figure 19 This is a schematic diagram of chip surface inspection during an inspection method based on complex chip type IPM according to an embodiment of the present invention. Figure 3 .

[0082] Figure 20 This is a schematic diagram of chip surface inspection during an inspection method based on complex chip type IPM according to an embodiment of the present invention. Figure 4 .

[0083] Figure 21 This is a schematic diagram of gold wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 1 .

[0084] Figure 22 This is a schematic diagram of gold wire detection during the detection of a detection method based on complex chip type IPM according to an embodiment of the present invention. Figure 2 . Detailed Implementation

[0085] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, further illustrating the present invention.

[0086] like Figure 1 As shown, the product composition includes: fine aluminum wires, gold wires, thick aluminum wires, three MOSFETs on the right, one IC in the middle, six large MOSFETs and six small MOSFETs below. The gold wires, fine aluminum wires, IC, and three MOSFETs on the right are 1-2mm higher than the middle row of fine aluminum wires, six large MOSFETs, six small MOSFETs, and six thick aluminum wires below. With a lens depth of field of 0.5mm, it is impossible to capture all the chips and bonding wires at once. Therefore, it is necessary to switch between different lights and heights to take multiple pictures to achieve a clear image of all chips and features.

[0087] First, combine Figures 1-22 This invention describes a detection method based on complex chip type IPM, which is used for testing IPM products and has a wide range of applications.

[0088] like Figures 1-22 As shown, an embodiment of the present invention provides a detection method for complex chip type IPM, which includes the following steps:

[0089] In S1, the bright and dark image acquisition: by controlling the switchable combination light source and adjusting the height of the lens, a clear image of all chips and bonding wires on the product is acquired, and the image is registered as a template.

[0090] Furthermore, the specific steps for obtaining the light and dark image are as follows:

[0091] S101: To photograph the first type of chip, turn on the coaxial light in the combined light source and move the lens to a height where the first type of chip can be clearly photographed.

[0092] S102: Adjust the type and brightness of the light source so that the chip in the image can be clearly imaged and the gray value of the chip surface is within the preset range. The preset range of gray value is about 100.

[0093] S103: Configure the currently adjusted chip image and register it as a template.

[0094] S104: Repeat the above process (i.e., S101, S102 and S103) until all chips that need to be photographed have completed image acquisition.

[0095] In S2, chip localization: Based on templates, a template matching algorithm is used to locate chips and frames in the current image, obtaining a transformation matrix representing their position and pose. Specifically, during template matching, a template region for the chip needs to be defined, which involves selecting the unique contours of the chip surface for model training. The transformation matrix obtained from the current chip localization is then passed out for use in subsequent algorithms. This process is repeated for all chips and frames. The transformation matrix is ​​as follows:

[0096] ;

[0097] Where a is the scaling component in the X-axis direction; d is the scaling component in the Y-axis direction; b and c control rotation and shearing; tx is the translation in the X-axis direction; and ty is the translation in the Y-axis direction.

[0098] In S3, wire bonding inspection: the inspection areas for aluminum and gold wires are determined, and the aluminum and gold wires are extracted within the inspection areas using image processing algorithms. Connectivity analysis is then performed to determine whether there are any breakage defects in the aluminum and gold wires.

[0099] Furthermore, the specific steps for aluminum wire testing are as follows:

[0100] In S311, a detection area is selected at the detection location. The coordinates of this selection area are then transformed using a transformation matrix to obtain the precise coordinates of the aluminum wire detection area. Specifically: First, a detection area needs to be selected at the detection location. The original coordinates (x, y) of this selection area are then affinely transformed based on the affine matrix obtained from the positioning to obtain the precise aluminum wire detection area. The formula is as follows:

[0101] ;

[0102] In S312, aluminum wires and aluminum wire solder joints in the detection frame are identified and extracted using a threshold method (the aluminum wires extracted by the threshold method are regions with specific shapes). The extracted aluminum wire regions are subjected to curvature analysis. If the curvature analysis meets the requirements, connectivity analysis is performed to determine whether the aluminum wire is broken.

[0103] Furthermore, in this embodiment, the extraction using the threshold method specifically involves:

[0104] Based on the current grayscale value f(x,y) of the aluminum wire, select its corresponding grayscale threshold T;

[0105] Specify whether the aluminum wire is black or white. For black aluminum wire, specify the grayscale range as 0~T. For white aluminum wire, specify the grayscale range as T~255.

[0106] ;

[0107] Furthermore, in this embodiment, the connectivity analysis specifically includes:

[0108] The extracted aluminum wire region is divided into multiple connected regions R1, R2, ..., Rn, satisfying...

[0109] ;

[0110] in, : Connect all connected regions R K The union of K=1 to n represents the foreground region of the entire image; R K The Kth connected region; : is the set of foreground pixels, that is, the set of all pixel coordinates with a value of 1; Any two distinct connected regions do not overlap;

[0111] If you find an area that runs through the first and second solder joints of the aluminum wire, it proves that the aluminum wire is not broken; otherwise, the aluminum wire is broken.

[0112] The aluminum wire detection steps described above are used to extract and analyze all aluminum wires on the current product.

[0113] Regarding the inspection of gold wire, it should be noted that the gold wire in the current product consists of a regularly shaped circular first solder joint, an irregularly shaped second solder joint, and the line connecting the two solder joints. When drawing the inspection frame, the inspection area needs to be accurately drawn for the first solder joint, the second solder joint, and the line.

[0114] Furthermore, in this embodiment, the specific steps for gold wire detection are as follows:

[0115] In S321, the detection of a single solder joint of gold wire is performed as follows: within the detection area of ​​a single solder joint, the center and radius of the circular solder joint are located by extracting the edge of a circumferential caliper and fitting a circle using the least squares method.

[0116] In S322, the detection of two solder joints of gold wire is as follows: within the detection area of ​​two solder joints, the irregular positions of two solder joints are located by grayscale template matching algorithm;

[0117] In S323, gold wire extraction: Based on the located positions of the first and second solder joints, a gold wire detection area is generated. Within this area, the gold wire region is extracted using a threshold method, and connectivity analysis is performed to determine whether the gold wire is broken. The gold wire detection process here is the same as the aluminum wire detection, so it will not be described further.

[0118] Furthermore, in this embodiment, the detection of a solder joint of the gold wire specifically involves:

[0119] Let the parameter of an initial circle be the center (x). c ,y c With radius r, N calipers are evenly placed on the circumference of the circle, centered at the circle's center. The coordinates of the center point of each caliper are:

[0120] ;

[0121] in, ; This is the initial angle;

[0122] Obtain the grayscale profile g on each caliper line segment. i (s), and find its first derivative (gradient):

[0123] ;

[0124] This indicates the location where the grayscale change is most drastic, i.e., the edge point;

[0125] The coordinates of this edge point in the image are:

[0126] ;

[0127] All edge points Fit to an optimal circle:

[0128] ;

[0129] The final center and radius of the circle are obtained by minimizing the error function:

[0130] .

[0131] Furthermore, in this embodiment, the detection of the two solder joints of the gold wire is specifically as follows:

[0132] Using a grayscale template matching algorithm, one or more standard two-weld-point templates are created, and weld points similar to the templates are searched in the two-weld-point detection area.

[0133] Gray-scale matching correlation coefficient:

[0134] ;

[0135] Where T: template grayscale set; The set of grayscale values ​​in the target image corresponding to the template position; Gray-scale dot product; : The grayscale energy product of the template and the target region;

[0136] Find the point with the maximum value as the matching result:

[0137] .

[0138] Through the above gold thread detection steps, all gold threads on the current product are extracted and analyzed.

[0139] In S4, chip surface inspection: It receives aluminum and gold wire area information from the wire bonding inspection step, shields the areas on the chip surface marked with lines, obtains the true surface inspection area, extracts chip surface defects through threshold segmentation, and distinguishes between dirt and scratches based on the shape characteristics of the defects. If the area is greater than a set value, it is considered dirt. If the aspect ratio is greater than a set value, it is considered a scratch. It should be noted that the transmission of aluminum and gold wire area information requires the use of a flowchart for connection.

[0140] Furthermore, in the chip surface inspection step, when detecting chipped edges, it is necessary to shrink the chip surface according to the outline of the inspection, and extract and determine the defect threshold within this loop-shaped area.

[0141] In S5, position detection: based on the position information obtained in the chip positioning step, calculate the chip's position offset and rotation angle relative to the frame, and determine whether it is within the tolerance range.

[0142] Furthermore, the specific steps for location detection are as follows:

[0143] S501: Based on the positioning information of the chip and the frame, the settings are made, with the frame as the reference system and the chip as the object under test.

[0144] S502: Offset range setting, the chip offset relative to the frame can be set to limit the range by setting the upward, downward, left and right offset of the chip relative to the frame according to the actual detection requirements;

[0145] S503: Angle range setting, the angle of the chip relative to the frame can be set to set the limit range according to the actual detection requirements.

[0146] In S6, a comprehensive judgment is made based on the results of the wire bonding inspection step, the chip surface inspection step, and the position inspection step. If all the inspection results are within the preset tolerance range, the product is deemed qualified; otherwise, the product is deemed unqualified.

[0147] Regarding the appendix Figures 3-22 The explanation is as follows:

[0148] Appendix Figures 3-7 This is a schematic diagram of obtaining multiple sets of light and dark images of an IPM product.

[0149] Appendix Figure 8 This is a schematic diagram showing the test results of each module on the IPM product.

[0150] Appendix Figures 9-12 The diagrams, in order, are: the middle row of chips shielded based on the location area of ​​the aluminum wire acquisition; the top chip shielded based on the location area of ​​the gold wire acquisition; the bottom row of chips shielded based on the location area of ​​the aluminum wire acquisition; and the three chips on the right shielded based on the location area of ​​the gold wire acquisition. (Attached) Figures 9-12 The red box area represents the shielded area.

[0151] Appendix Figures 13-16 The image shows an enlarged schematic of aluminum wire inspection, where the green box represents the aluminum wire area being inspected, and the yellow box represents the first and second solder joint areas being inspected.

[0152] Appendix Figures 17-20 Here is an enlarged schematic diagram of chip surface inspection, where the red box represents the detected surface contamination area.

[0153] Appendix Figures 21-22 The diagram shows an enlarged view of the gold wire inspection, where the green box represents the gold wire area being inspected, and the yellow box represents the first and second solder joint areas being inspected.

[0154] Above, refer to Figures 1-22 A detection method based on complex chip type IPM according to an embodiment of the present invention is described, which has the following beneficial effects:

[0155] 1. This IPM detection method is a general detection process that combines multiple sub-algorithms, is adapted to the characteristics of IPM products, and has strong versatility.

[0156] 2. Solve the problem of unclear imaging of all chips and bonding wires due to large height differences on the product surface.

[0157] 3. The testing process is flexible and can be tailored to different types of chips.

[0158] 4. It can adapt to different numbers of chip inspections and different types of wire bonding inspections. Even if more types are added, the inspection process and algorithm can still perform the inspections.

[0159] 5. Before chip testing, the accurate positions of gold and aluminum wires can be obtained for shielding, improving the accuracy of chip surface testing.

[0160] It should be noted that, in this specification, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0161] Although the present invention has been described in detail through the preferred embodiments above, it should be understood that the above description should not be considered as a limitation of the present invention. Various modifications and substitutions to the present invention will be apparent to those skilled in the art after reading the above description. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims

1. A method for detecting a complex chip type IPM, characterized in that, The method comprises the following steps: Bright-dark image acquisition: by controlling the switchable combined light source and adjusting the height of the lens, clear images of all chips and bonding wires on the product are acquired, and the images are registered as templates; Chip positioning: based on the templates, the chips and frames in the current image are positioned by a template matching algorithm to obtain a transformation matrix representing the position and posture; Bonding wire detection: the aluminum wire and gold wire detection regions are determined, the aluminum wire and gold wire in the bonding wire detection region are extracted by an image processing algorithm, and connectivity analysis is performed to determine whether the aluminum wire and gold wire are broken; Chip surface detection: receiving the aluminum wire and gold wire region information output by the bonding wire detection step, shielding the chip surface using the line region, obtaining the real surface detection region, extracting the chip surface defects by threshold segmentation, and distinguishing dirt and scratches according to the shape characteristics of the defects; Position detection: based on the position information obtained in the chip positioning step, the position offset and rotation angle of the chip relative to the frame are calculated, and it is determined whether it is within the tolerance range; Comprehensive judgment: the results of the bonding wire detection step, the chip surface detection step and the position detection step are comprehensively judged, if all the detection results are within the preset tolerance range, the product is determined to be qualified, otherwise the product is determined to be unqualified.

2. The detection method based on complex chip type IPM as described in claim 1, characterized in that, The bright-dark image acquisition step specifically comprises: Photographing the first type of chip, turning on the coaxial light in the combined light source, and moving the lens to a height where the first type of chip can be clearly photographed; Adjust the type and brightness of the light source so that the chip in the image can be clearly imaged, and the chip surface gray value is within the preset range; Set the current adjusted chip image as a template; Repeat the above process until all the chips that need to be photographed have completed image acquisition.

3. The detection method based on IPM for complex chip types as described in claim 1, characterized in that, The specific steps of aluminum wire detection are: Frame the detection region for the detected position, transform the detection frame coordinates by the transformation matrix to obtain accurate aluminum wire detection region coordinates; Identify the aluminum wire and aluminum wire pads in the detection frame, extract them using threshold method, and analyze the curvature of the extracted aluminum wire region. If the curvature analysis meets the requirements, connectivity analysis is performed to determine whether the aluminum wire is broken.

4. The detection method based on IPM for complex chip types as described in claim 3, characterized in that, The threshold method for extraction specifically comprises: According to the gray value f(x, y) of the current aluminum wire, select its gray threshold T; Determine whether the aluminum wire is black or white. The gray interval of black aluminum wire is 0~T, and the gray interval of white aluminum wire is T~255.

5. The detection method based on IPM for complex chip types as described in claim 3, characterized in that, The connectivity analysis specifically comprises: The extracted aluminum wire region is divided into multiple connected regions R1, R2,..., Rn, which satisfy ； wherein, : all connected regions R K : union from K=1 to n, representing the foreground region of the whole image; R K : the Kth connected region; : is the set of foreground pixels, i.e., the set of pixel coordinates with all values of 1; : any two different connected regions do not overlap with each other; Find a region that can pass through the aluminum wire one and two pads, then prove that the aluminum wire is not broken, otherwise the aluminum wire is broken.

6. The method of claim 1, wherein the complex chip type IPM is detected based on a complex chip type IPM detection method, and the complex chip type IPM detection method comprises: The specific steps of gold wire detection are: ​ One pad detection of gold wire: in the one pad detection region, the center and radius of the circular one pad are positioned by circular caliper edge extraction and least squares circle fitting; Two pad detection of gold wire: in the two pad detection region, the position of irregular two pads is positioned by gray template matching algorithm; Gold wire extraction: according to the position of the first and second welding points, a gold wire detection area is generated, the area of the gold wire is extracted in the area by threshold method, and connectivity analysis is performed to determine whether the gold wire is broken.

7. The detection method based on IPM for complex chip types as described in claim 6, characterized in that, The first welding point detection of the gold wire is specifically: Let the parameters of an initial circle be the center (x c ,y c ) and the radius r. N calipers are evenly arranged on the circumference with the center as the center, and the coordinate of the center point of each caliper is ； wherein ; is the initial angle; On each caliper line segment, the gray scale profile g is taken i (s) and its first derivative is taken: ； indicates the position where the gray scale changes most drastically, i.e. the edge point; The coordinate of the edge point in the image is: ； All edge points ( , Fit an optimal circle: ； The final center and radius are obtained by minimizing the error function: 。 8. The detection method based on IPM for complex chip types as described in claim 6, characterized in that, The second welding point detection of the gold wire is specifically: A standard second welding point template or multiple standard second welding point templates are established using a gray scale template matching algorithm, and similar welding points are searched in the second welding point detection area. Gray scale matching correlation coefficient: ； Wherein, T: template gray set; : Target image corresponding to the template position of the gray value set; : Gray point multiplication; : Template and target area of gray energy product; The maximum point is found as the matching result: 。 9. The detection method based on IPM for complex chip types as described in claim 1, characterized in that, In the chip surface detection step, when detecting edge collapse, the profile of the chip surface detection is needed to be retracted, and the defect threshold extraction and determination are performed in the retracted area.

10. The detection method based on IPM for complex chip types as described in claim 1, characterized in that, The specific of the position detection step is: According to the positioning position information of the chip and the frame, the frame is set as the reference system, and the chip is selected as the measured object; Offset range setting: according to the actual detection requirement, the offset of the chip relative to the frame is set to set the limit range; Corner range setting: according to the actual detection requirement, the corner of the chip relative to the frame is set to set the limit range.

Images