Appearance inspection device

Abstract

The appearance inspection device (10) provides a highly reproducible and accurate appearance inspection technique. The appearance inspection device (10) is characterized by comprising: a camera (32) consisting of an imaging optical system and an imaging element, which acquires an image (18) used in the appearance inspection; a mechanism (31) that moves a distance equal to the division distance when the field of view of the camera (32) is divided into M×N (2≤M×N) equal parts, thereby changing the positional relationship between the camera and the object under inspection (30); and an image processing inspection unit (23) that processes and inspects the image (18) obtained from the camera (32), and acquires at least one image (18) by the camera (32) each time the mechanism (31) moves a distance, acquiring M×N images of the entire inspection area of ​​the object under inspection (30), and performing inspection processing on the acquired images (18) by the image processing inspection unit (23).

Description

Technical Field

[0001] This invention relates to visual inspection technology that utilizes image processing technology to inspect various products. Background Technology

[0002] Visual inspection devices that use cameras as image acquisition equipment to perform visual inspections of various products are becoming increasingly common. A camera consists of an imaging optical system that forms an optical image of the inspected object and an imaging element (usually a CMOS image sensor) that photoelectrically converts the light intensity of the image to form an electronic image. The size of the image obtained in a single shot is called the field of view.

[0003] While the field of view can vary depending on the lens magnification, in inspection, the size of the object being inspected (equivalent to one pixel of the imaging element), i.e., the pixel resolution, is the primary factor determining inspection performance. Therefore, the magnification of the imaging optical system must be determined to obtain the desired pixel resolution, and the field of view cannot necessarily be freely set. Thus, when inspecting objects larger than the field of view, the camera or object is moved, and multiple fields of view are used to inspect the entire object. For example, with a camera field of view of 11×11mm and an object size of 20×20mm, using four fields of view allows for the inspection of the entire object. This type of inspection will be referred to as multi-field inspection.

[0004] As a concrete example, the following describes bump inspection in the semiconductor packaging process (back-end process). IC (integrated circuit) chips require resin encapsulation to cope with degradation and heat. Packaging involves mounting the chip on a multilayer wiring substrate, connecting the electrodes on the multilayer wiring substrate to the electrodes on the chip, and then sealing it with resin. However, traditionally, the connection between electrodes was generally achieved using wire bonding. In recent years, however, flip-chip mounting, which mainly uses a two-dimensional array of spherical micro-electrodes (hereinafter referred to as "bumps") for connection, has become increasingly popular.

[0005] In flip-chip mounting, where electrodes are directly aligned, consistent bump size becomes crucial. This is because any deviation in bump size can lead to poor connections. Therefore, the bump morphology needs to be accurately checked before connection (see, for example, Patent Document 1).

[0006] Visual inspection of bumps, as part of the online inspection process in manufacturing, is performed by photographing the bumps arranged in an array from the top surface. In recent years, 100% inspection has become common, making inspection speed crucial. Currently, bumps range in diameter from 10 to 80 μm, arranged in hundreds to tens of thousands at a spacing of 20 to 130 μm on each multilayer wiring board. Furthermore, it is required to inspect all these bumps at a high speed of 1 to tens of seconds per multilayer wiring board.

[0007] In recent years, packaging using the concept of chip-on-chip has been promoted, which mounts multiple chips on a multilayer wiring board. As a result, the size of the multilayer wiring board has become larger. On the other hand, due to the miniaturization of circuits, the diameter of the bumps has become smaller. Fine pixel resolution is required in appearance inspection, so the field of view has become smaller. As a result, multi-field inspection using a large number of fields of view has become necessary.

[0008] Existing technical documents Patent documents Patent Document 1: Japanese Patent No. 4984332 Summary of the Invention

[0009] The problem that the invention aims to solve As a problem point in image-based inspections, there are issues with imaging optical system aberrations. Due to these aberrations, image quality is uneven within the field of view; for example, the aberration state differs between the center and periphery of the field of view, resulting in different image quality. If the image quality differs, the inspection results may also differ. For instance, in checking the height of a protrusion, even for the same protrusion, the height value at the center of the field of view may differ from the height value at the periphery. (First problem) Furthermore, if the inspected object becomes miniaturized and high-precision inspection is required, the inspection results may be affected by the pixel configuration of the digital camera and the positional relationship of the inspected object. For example, a spherical bump is relatively close to a mirror surface; therefore, when illuminated from directly above, the reflected light is limited to the vertex portion. Thus, as the bump diameter decreases, the area where reflected light returns also decreases, and the size of the captured vertex portion becomes smaller than the camera's pixel size. Consequently, there are cases where the reflected light from the vertex portion of the bump is captured using a single pixel, and cases where it is captured using multiple adjacent pixels. Therefore, depending on which case, the morphological information of the bump differs, leading to a decrease in the accuracy and reproducibility of the inspection results (Second Problem).

[0010] This invention was made with this in mind, and its purpose is to provide a highly reproducible and accurate appearance inspection technique.

[0011] Methods for solving problems The appearance inspection apparatus of the present invention is characterized by comprising: a camera, which is composed of an imaging optical system and an imaging element, for acquiring images for appearance inspection; a mechanism that changes the positional relationship between the camera and the object under inspection by a moving distance approximately equal to the dividing distance when the field of view of the camera is divided into M×N (2≤M×N) equal parts; and an image processing inspection unit that processes and inspects the images obtained from the camera, and acquires at least one image by the camera each time the moving distance is moved by the mechanism, acquiring M×N images of the entire inspected area of ​​the object under inspection, and performing inspection processing on the acquired images by the image processing inspection unit.

[0012] Invention Effects According to the present invention, a more accurate appearance inspection technology is provided. Attached Figure Description

[0013] Figure 1 This is a block diagram of an appearance inspection device according to the first embodiment of the present invention.

[0014] Figure 2 (A) is a top perspective view of a multilayer wiring substrate with bumps arranged in an array. Figure 2 (B) is a side perspective view of a multilayer wiring substrate on which chips are mounted via bumps.

[0015] Figure 3 (A) ~ Figure 3 (F) is an operational illustration of the appearance inspection device according to the embodiment.

[0016] Figure 4 (G)~ Figure 4 (L) is an operational illustration of the appearance inspection device in the embodiment.

[0017] Figure 5 (A) is a side view showing the reflected light from the apex incident on the camera through convex points of different sizes. Figure 5 (B) is its top view.

[0018] Figure 6 (A) represents an image of a convex point where a single pixel detects reflected light from a vertex. Figure 6 (B) represents an image of a convex point where reflected light from a vertex is detected by two pixels. Figure 6 (C) represents the image of a convex point where reflected light from a vertex is detected by four pixels.

[0019] Figure 7 It is an expression related to the algorithm of appearance inspection. Detailed Implementation

[0020] (First Implementation) Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. For specific explanation, the following description will use a bump inspection as an example. Figure 1 This is a block diagram of the appearance inspection device 10 according to the first embodiment of the present invention. Figure 2 (A) is a top perspective view of a multilayer wiring substrate (examined object) 30 in which the bumps 14 are arranged in an array. Figure 2 (B) is a side perspective view of IC package 37, in which chip 36 is mounted on multilayer wiring substrate 30 via bumps 14.

[0021] like Figure 2 As shown, in the IC package 37, the electrodes of the multilayer wiring substrate 30 (hereinafter referred to as "substrate 30") and the electrodes of the IC chip 36 are electrically connected via bumps 14. This method of electrical connection using bumps 14 is called flip-chip mounting. In this embodiment, the bumps 14 arranged on the substrate 30 are inspected. Then, the chip 36 is mounted, and further, to cope with degradation and heat, it is encapsulated with resin 35, completing the IC package 37. Furthermore, the substrate 30 with arranged bumps 14 is exemplified as the object of inspection for the visual inspection apparatus 10 in this embodiment, but it is not particularly limited.

[0022] like Figure 1 As shown, there are 14 protrusions arranged in a row. Figure 2 Multiple substrates 30 (shown as 3×5=15 in the illustration) are housed on a tray 38. The tray 38, containing multiple substrates 30, is placed on an inspection table 39 for inspection by the visual inspection device 10. After inspection, the tray 38 is moved to the next process, where the tray 38 containing uninspected substrates 30 is placed back on the inspection table 39. Furthermore, a multilayer wiring substrate is shown as an example of the substrate 30 with arranged bumps 14, but there is no particular limitation.

[0023] Figure 3 (A)~(F) and Figure 4 (G)~(L) are operation illustrations of the appearance inspection device 10 with protrusion 14 in the embodiment. Also refer to Figure 1 Continuing the explanation, the field of view 12 of the camera 32 is divided into four equal parts of 2×2, each referred to as segment 11. The system includes: a mechanism 31 that moves the inspected area 14a, which has a set of protrusions 14, sequentially through the field of view 12 in a stepped manner with a movement distance equal to the segment spacing of the segments 11; and an image processing and inspection unit 23 that performs image processing and inspection on the image 18 of the field of view 12 acquired from the camera 32.

[0024] Furthermore, in the above explanation, section 11 is defined as the size that divides the field of view 12 of camera 32 into four equal parts, but this does not need to be an exact size. When performing multi-field inspection, some overlapping areas are generally defined between adjacent fields of view. For example, if protrusions are considered, if protrusions are divided into adjacent fields of view and photographed, the measurement process becomes complicated. Therefore, in order to photograph without necessarily dividing them into arbitrary fields of view, overlapping photographs are taken in an overlapping manner, such that areas greater than the distance between protrusions are photographed in any field of view. Except for such areas, dividing into four equal parts is sufficient; it does not need to be strictly 1 / 4 of the field of view. Figure 3 and Figure 4 The situation is indicated by the inconsistency between the dimensions of the field of view 12 and the four segments 11.

[0025] Furthermore, dividing into four equal parts is one example, which is generalized to M×N equal parts (2≤M×N). Thus, by moving the image to the field of view 12 in a stepwise manner with a distance equal to the segmentation interval of section 11, all the protrusions are photographed M×N times. Therefore, the morphological information 24 of the protrusion 14 can be determined using the image information from M×N times, thereby improving the determination accuracy.

[0026] Mechanism 31 feeds the cameras 32 sequentially in a stepped manner with a movement distance equal to the dividing distance of section 11. Furthermore, as a variation of mechanism 31, although not shown in the figure, it is also possible to sequentially transport the inspection table 39, which carries the tray 38, in a stepped manner.

[0027] Regarding the operation of the protrusion visual inspection device 10, firstly as follows: Figure 3 As shown in (A), the field of view 12 of the camera 32 is positioned in a manner that includes a segment 11 located at the corner of the inspected area 14a where a set of protrusions 14 exist, in the image processing inspection unit 23. Figure 1 Image 18 was obtained from ).

[0028] Next, as Figure 3 As shown in (B), the field of view 12 is advanced one step along the arrangement of the protrusions 14 with a movement distance equal to the segmentation spacing, and the image 18 is acquired in the image processing inspection unit 23. Furthermore, as... Figure 3 As shown in (C), the field of view 12 is fed sequentially and the image 18 is acquired repeatedly along the arrangement of the protrusions 14 until the opposite end. Then, it is fed one step in the right-angle direction, as shown in (C). Figure 3 As shown in (D), the sequential feeding of the field of view 12 and the acquisition of image 18 are repeated in the opposite direction until... Figure 3 Up to the opposite end shown in (E). Similarly, from... Figure 3 (F) to Figure 4The (G)~(L) are repeatedly fed in a step-like manner with a moving distance equal to the segmentation distance of segment 11 and the image 18 is acquired.

[0029] In this way, the image processing inspection unit 23 acquires M×N (4 in the illustration) images 18 for each of the multiple segments 11. Furthermore, the same segment 11 captured in the multiple images 18 were taken in different regions of the M×N segmented field of view 12.

[0030] Typically, image 18 is affected by aberrations in the imaging optical system. Generally, aberrations increase from the center of the field of view towards the periphery, tending to degrade image quality. Moreover, the effect of these aberrations is not necessarily point-symmetrical due to manufacturing errors such as eccentricity. On the other hand, the protrusions 14 captured in different regions of the field of view 12 are located at locations where the aberration conditions of image 18 differ, and are considered to have different levels of accuracy. Therefore, the M×N images 18 (4 in the illustration) are either images detected at the center of the field of view 12 or images detected at the periphery. Furthermore, by statistically processing these detection values, the accuracy of determining the morphological information 24 of the protrusions 14 can be made consistent.

[0031] Figure 5 (A) is a side view showing reflected light 26 (26x, 26y) from vertex 27 incident on camera 32 at protrusions 14 (14x, 14y) of different sizes. Figure 5 (B) is its top view. The spherical protrusion 14 is relatively close to the mirror surface, so when illuminated from directly above, the reflected light 26 (26x, 26y) incident on the camera 32 is confined to a portion of vertex 27 (27x, 27y). Therefore, as the diameter of the protrusion 14 decreases, the area of ​​vertex 27 to which the reflected light 26 returns also decreases.

[0032] Figure 6 (A) is an image 18 representing the case where a single pixel 22 detects reflected light 26 from vertex 27. Figure 6 (B) is an image 18 showing the case where the reflected light 26 from vertex 27 is detected using two pixels 22. Figure 6 (C) is an image 18 showing the case where the reflected light 26 from the vertex 27 is detected using four pixels 22. Thus, when the convex point 14 becomes smaller and the size of the captured vertex 27 is less than the size of the camera's pixel 22, there are cases where the vertex 27 is captured with one pixel and cases where the vertex 27 is captured with multiple adjacent pixels.

[0033] The movement distance of the field of view 12 is the size of the segment 11 (=segmentation distance). However, by setting the movement distance to an integer multiple of the pixel resolution and adding 1 / M (for vertical movement distance) or 1 / N (for horizontal movement distance) of the pixel resolution, if the movement amount is accurate, the M×N image 18 can obtain images with the positional relationship of the convex point relative to the pixel 22 offset by 1 / M vertically and 1 / N horizontally.

[0034] That is, the value obtained by multiplying the segment spacing of section 11 by the magnification of the imaging optical system and dividing it by the size of the pixel 22 of the camera 32 (pixel size) is rounded to an integer value close to the value. Then, 1 / M times the pixel size is added vertically and 1 / N times horizontally. The values ​​obtained by dividing these values ​​by the magnification of the imaging optical system become the vertical and horizontal dimensions of the movement distance.

[0035] This pixel shift confirms the effect of increased pixel resolution, which can improve... Figure 6 The difference in the positional relationship between the pixels and the object shown can affect the measured values, thereby improving the accuracy and reproducibility of the inspection results.

[0036] Image processing inspection unit 23 ( Figure 1 The shape information 24 of the bump 14 of interest is determined based on the four (M×N) vertex pixels 22a obtained by sequential feeding. Specifically, the shape information 24 includes the position, height, outer diameter of the bump 14, shape anomalies of other bumps 14, height of the substrate 30, surface foreign objects, etc.

[0037] The measurements of this morphological information 24 are obtained based on the brightness values ​​of each of the four (M×N) images 18 representing the various shapes of the convex points 14, with one value calculated for each of the vertex pixel 22a and / or its surrounding pixels 22, totaling four (M×N). These multiple measurements are then statistically processed to output a representative value for each convex point 14. Furthermore, the measurements and representative values ​​of the morphological information 24 representing the convex points 14 are not limited to a single item, but are determined for multiple items.

[0038] Figure 7 This refers to the formula related to the algorithm for appearance inspection. The representative value V of morphological information 24 is calculated based on multiple measurements v according to several (3) algorithms exemplified below. i However, the calculation method for the value V is not limited to these algorithms.

[0039] The first algorithm is the "simple average". This "simple average" is based on... Figure 7 Equation (1) applies to M×N measured values ​​v iThe method of averaging (i=1~M×N) to calculate the representative value V. Based on the effects of averaging (law of large numbers, central limit theorem), and choosing one measurement value v... i Compared to the representative value V, reliability is improved.

[0040] The second algorithm is a "reliability-based weighted average". This "reliability-based weighted average" is based on... Figure 7 Equation (2) will derive M×N measurement values ​​v respectively. i The reliability value t of pixel 22 (i=1~M×N) i The weighting factor is multiplied by the corresponding measured value v. i Then, a weighted average is performed to calculate the representative value V. This "reliability-based weighted average" can suppress... Figure 6 The pixel segmentation effect shown in (B) and (C) represents an improvement in the reliability of the value V.

[0041] Here, the reliability value is a quantity that represents the degree of reliability of the inspection or measurement values ​​of each pixel, and it varies depending on the method of visual inspection. For example, the brightness value of a pixel represents the amount of light as a signal, so it can be said to represent reliability in the sense of whether the signal is reliably obtained, and is an example of a reliability value.

[0042] The third algorithm is a "weighted average based on distance from the center". This "weighted average based on distance from the center" is calculated according to... Figure 7 Equation (3) will derive M×N measurement values ​​v respectively. i The distance L from pixel 22 (i=1~M×N) to the center of the field of view 12 i (i=1~M×N) is used as a weighting coefficient multiplied by the corresponding measured value v i Then, a weighted average is performed to calculate the representative value V. This "weighted average based on distance from the center" emphasizes values ​​measured near the center within the field of view, thereby suppressing the influence of aberrations and improving the reliability of the representative value V.

[0043] According to the present invention, the influence of aberrations in the imaging optical system (first issue) and pixel fragmentation (second issue) in visual inspection using images can be reduced, thereby improving the reliability of the inspection. On the other hand, there is a negative consequence of a significant increase in inspection time. The larger M×N is, the higher the reliability of the inspection, but the inspection time increases. However, it is considered a great advantage that by incorporating the functions of the present invention into the visual inspection device, the user can select the optimal reliability and speed.

[0044] Explanation of reference numerals in the attached figures 10…Visual inspection device 11… section 12… field of view 14…bumps 14a…A set of protrusions (inspected area) 15… Adjustment Department 18…images 22… pixels 22a…vertex pixels 23…Image Processing Inspection Department 24…morphological information 26…Reflected light 27…Vertex 30… Multilayer wiring substrate (substrate, object under inspection) 31…Institutional Department 32… cameras 35… package 36…IC chips 37…IC package 38… trays 39… Inspection Table

Claims

1. A visual inspection device, characterized in that, have: A camera, consisting of an imaging optical system and a camera element, acquires images used in visual inspection; The mechanism changes the positional relationship between the camera and the object under inspection by a moving distance equal to the dividing distance when the field of view of the camera is divided into M×N equal parts, where 2≤M×N; as well as The image processing and inspection unit processes and inspects the images obtained from the camera. Each time the mechanism moves the distance, at least one image is acquired by the camera, and all M×N images of the inspected area of ​​the object are acquired. The acquired images are then inspected by the image processing and inspection unit.

2. The appearance inspection device according to claim 1, characterized in that, The value obtained by multiplying the segmentation distance by the magnification of the imaging optical system and dividing it by the pixel size of the camera is rounded to an integer value close to the value. Then, 1 / M times the pixel size is added vertically and 1 / N times the pixel size is added horizontally. The values ​​obtained by dividing these values ​​by the magnification of the imaging optical system are set as the vertical and horizontal dimensions of the movement distance.

Images