CN108596875A - A kind of semiconductor chip flash rapid detection method based on image segmentation algorithm - Google Patents

Abstract

The invention discloses a semiconductor chip overflow rapid detection method based on an image segmentation algorithm, aiming at identifying the excess spilled onto pins and exposed carriers during the EMC (Epoxy Molding Comp) packaging process of an integrated circuit, so as to provide EMC packaging The process provides timely feedback information. First, preprocess the original image of the EMC package collected by the industrial camera, and then use the Selective search algorithm to segment the image to obtain the target candidate area, and screen out the area of the chip to be detected, and then filter the selected area. The area of the threshold is segmented to obtain the overflow area. This method can give the flash detection result of each chip in a very short time, which not only optimizes the flash detection method in the EMC packaging process, but also provides real-time feedback information for the EMC packaging process, thereby improving EMC packaging. performance.

Description

A Fast Detection Method for Semiconductor Chip Flashing Based on Image Segmentation Algorithm

technical field

The invention belongs to the field of integrated circuit EMC packaging overflow detection, in particular to a method for quickly detecting semiconductor chip overflow based on an image segmentation algorithm.

Background technique

EMC is an important microelectronic packaging material. Through the packaging process, the semiconductor chip is covered to form a protection, so as not to be damaged by the external environment. At the same time, EMC also has a certain heat dissipation effect. EMC has the characteristics of large-scale production and reasonable reliability. It is one of the common packaging materials for semiconductor packaging and occupies more than 95% of the packaging market with its unique advantages. However, in the semiconductor plastic packaging process, due to the gap between the lead frame and the plastic packaging film, the gel-like plastic packaging material will leak out from the gap, solidify on the surface of the lead body, and form black flashes, which are integrated circuit pins and exposed parts. excess on the carrier. The flash itself will not have adverse effects on the performance of the plastic packaged product, but because the overflowed plastic package covers the lead pins, if not handled, it will form an insulating area and affect the quality of subsequent electroplating and product reliability, resulting in product open circuit and virtual soldering And other issues. Therefore, how to reduce the occurrence of flash and how to remove it has always been an important issue that has plagued researchers and manufacturers.

The causes of EMC packaging flash are usually: 1) The molding compound has good fluidity and low viscosity, and the molding compound/epoxy resin may overflow from the parting surface and cover the lead pin when the mold is not tightly clamped; 2) After long-term use of the plastic packaging mold, the surface is worn or the base is uneven, resulting in overflow; 3) During the semiconductor packaging process, the precision of the frame does not match the mold, and insufficient meshing causes edge pressing problems. After the flash is formed, it has no effect on the performance of the molded product, but because the overflowed molded compound covers the lead pin, it may form a plating defect and affect the reliability of the product. At present, manufacturers generally use a file, sandpaper, and weak acid cleaning to remove the residual flash after the flash removal process. However, this method is easy to scratch the plastic package and cause the surface of the lead pin to be rough, which affects the appearance quality. And electroplating quality, but also affect the production schedule. Therefore, how to avoid and reduce the generation of overflow is particularly important. In summary, it is of great practical significance to develop a visual, green and pollution-free flashing detection method to quickly detect the chip flashing situation to meet the needs of providing reference for the EMC packaging process in a timely manner.

Thanks to the development of computer technology and image processing technology, a new type of visual flash detection technology has emerged. The detection result is a grayscale image, which has the characteristics of high detection accuracy, flexible use, and large amount of information. It is very suitable for chip packaging production Process spill detection.

Contents of the invention

The purpose of the present invention is to provide a rapid overflow detection strategy for the overflow problem existing in the existing EMC packaging process. This strategy can accurately evaluate the chip flashing situation on site, and then provide feedback information for the adjustment and maintenance of in-service equipment, which is an important reference for the EMC packaging process.

The object of the present invention is achieved by the following technical solutions: a semiconductor chip flash detection method based on an image segmentation algorithm, the method comprises the following steps:

A semiconductor chip flash detection method based on an image segmentation algorithm is characterized in that the method comprises the following steps:

(1) Use the semiconductor packaging press (FSAM120-1US) and QFNQFNWB7X7-48L (T0.75) mold to produce the original packaging film, and use the industrial camera to obtain the multi-frame grayscale image of the original packaging film;

(2) Gaussian blur preprocessing is performed on each frame of image;

(3) Based on the image segmentation of selective search, hundreds of candidate regions are obtained and the chip regions to be analyzed are screened out. This step is realized by the following sub-steps:

(3.1) Perform over-segmentation on a single-frame image, and obtain n initial regions r ₁ , r ₂ ,...,r _n by segmenting to form an initial region set R={r ₁ ,r ₂ ,...,r _n } ; At the same time, construct an initialization similarity set S,

(3.2) Calculate the similarity s(r _i , r _j ) of each two adjacent regions, the similarity index adopts color similarity, size similarity, texture similarity, matching similarity, four complementary similarities, respectively Consolidate regions from different perspectives;

(a) The calculation method of color similarity is as follows:

Divide each initial area into q intervals according to the pixel value, and obtain a normalized color histogram in, Represents the normalized frequency of the k-th pixel value interval in the i-th initial region;

The color similarity of two adjacent areas r _i , r _j is:

The color histogram of the new region r _ij obtained by merging r _i and r _j is:

size(r _ij )=size(r _i )+size(r _j ) (19)

Where size(r _i ) is the size of the initial region r _i ;

(b) The calculation method of texture similarity is:

For each initial region, calculate its SIFT (scale-invariant feature transform) feature, where the scale factor is set to 1 in this patent, and the width of the histogram interval is 10, resulting in obtained as texture features.

in, Indicates the k-th texture feature in the i-th initial region;

The texture similarity of two adjacent regions r _i and r _j is:

The texture histogram of the new region r _ij obtained by merging r _i and r _j is

(C) The calculation method of size similarity is:

Where size(im) represents the size of the entire grayscale image.

(d) The calculation method of matching similarity is:

where BB _ij is a bounding box that only encloses regions r _i and r _j .

Finally, the four similarities are combined to obtain the final similarity:

s(r _i ,r _j )＝a ₁ Scolor(r _i ,r _j )+a ₂ Texture(r _i ,r _j )+a ₃ Ssize(r _i ,r _j )+a ₄ Sfill(r _i ,r _j )(24)

where a _i ∈ {0,1}, indicates whether the corresponding similarity is adopted.

(3.3) All the similarities form a similarity set S, and the two adjacent areas with the largest similarity s(r _i , r _j )=max(S) in the similarity set, merge these two areas r _ij =r _i ∪r _j , calculate the similarity between the new area r _ij and the adjacent area, update the similarity set: delete all the similarities s(r _i ,r _* ) related to the merged area r _i , and All similarities s(r _j , r _# ) related to the merged region r _j , added to the updated similarity set is: add the similarity between the new region r _ij and the adjacent region;

(3.4) Repeat steps 3.2-3.3, when the similarity set S is an empty set, the loop ends. Complete the merging of the initial regions to obtain the candidate region set R;

(3.5) Select the upper and lower chip regions to be analyzed from the candidate region set R, so that the size of the region to be analyzed is within the range of one-third to one-half of the size of the original package, and the two regions to be analyzed need to be analyzed separately Include and For these 4 key geometric points, the resolution of the packaged original film is a*b.

(4) Carry out threshold-based image segmentation to the filtered region to be analyzed, this step is realized by the following sub-steps:

(4.1) Calculate the normalized histogram of the area to be analyzed. Let {0,1,...,L-1} represent L different gray levels in the area image with size M*N pixels, and n _z represent the number of pixels with gray level z. The total number of pixels in the region image is MN=n ₀ +n ₁ +n ₂ +...+n _L-1 , p _z =n _z /MN,z=0,1,2,...,L-1 Indicates the normalized frequency of the gray level z in the histogram.

(4.2) For u=0,1,2,...,L-1, calculate the cumulative sum P ₁ (u), the calculation formula is as follows:

(4.3) For u=0,1,2,...,L-1, calculate the cumulative mean value m(u), the calculation formula is as follows:

(4.4) Calculating the global gray mean value m _G , the calculation formula is as follows:

(4.5) For u=0,1,2,...,L-1, calculate the variance between classes Calculated as follows:

(4.6) obtain the Otsu threshold u ^* by making the objective function (13) obtain the maximum value, if the maximum value is not unique, obtain u ^* with the average of each maximum value u detected correspondingly;

(4.7) Segment the region image based on the optimal threshold u ^* using the following formula to obtain the segmented image:

f(x, y) is the pixel of the pixel point (x, y) in the image of the area to be analyzed. After the segmentation, the pixel point of 1 represents the detected overflow part in the production process.

The beneficial effects of the present invention are: the present invention is a fast flash detection method oriented to the EMC packaging process, which divides the image into hundreds to thousands of candidate areas through a selective search algorithm, and screens out the chip area to be analyzed from the candidate areas . Further use the optimal threshold to segment the image to be analyzed to obtain the flash image, so that the seriousness of the flash can be measured. For each frame of image, the present invention can provide detection results within 20 seconds, can quickly feed back flash information for the EMC packaging process, provide references, and avoid more flash situations.

Description of drawings

Fig. 1 (a) is the flow chart of semiconductor packaging of the present invention, Fig. 1 (b) is the parameter setting in the actual production operation, and Fig. 1 (c) is the final packaging original film;

Fig. 2 is a semiconductor packaging overflow detection system of the present invention, the original grayscale image is obtained by an industrial camera, and is transmitted to a computer for further analysis;

Fig. 3 is the encapsulation original film grayscale image that the present invention obtains;

Fig. 4 is the reduced-resolution image that the method of the present invention utilizes Gaussian blur preprocessing to obtain, and has retained image complete information while reducing image resolution;

Fig. 5 is the display of partial results obtained after the selective search of the image is segmented by the method of the present invention;

Fig. 6 is the area to be analyzed screened out by the method of the present invention;

Fig. 7 is the final flash image after threshold segmentation obtained by the method of the present invention.

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings and specific examples.

(1) EMC (Epoxy molding compound) packaging: As shown in Figure 1(a), place the lead frame in the mold so that each chip is located in the acupuncture point, and put the block EMC into the mold space after the mold is closed. Set the parameters of the packaging press as shown in Figure 1(b). At high temperature, the EMC begins to melt, and flows along the track to the acupuncture points, gradually covering the chip from the bottom until it is completely covered and packaged, and then molded and solidified, as shown in Figure 1(c) shown;

(2) Image acquisition and preprocessing: The image-based flash detection system is shown in Figure 2. The industrial camera is used to acquire the grayscale image of the original packaging film as shown in Figure 3, and its resolution is 3264*4896. Blur processing reduces its resolution to 408*612, while retaining the integrity of image information, it removes unnecessary noise, reduces image resolution and speeds up the speed of flash analysis and detection, as shown in Figure 4.

(3.1) Perform over-segmentation on a single-frame image, and obtain n initial regions r ₁ , r ₂ ,...,r _n by segmenting to form an initial region set R={r ₁ ,r ₂ ,...,r _n } ; At the same time, construct an initialization similarity set S,

(3.2) Calculate the similarity s(r _i , r _j ) of each two adjacent regions, the similarity index adopts color similarity, size similarity, texture similarity, matching similarity, four complementary similarities, respectively Consolidate regions from different perspectives;

(a) The calculation method of color similarity is as follows:

Divide each initial area into q intervals according to the pixel value, and obtain a normalized color histogram in, Represents the normalized frequency of the k-th pixel value interval in the i-th initial region;

The color similarity of two adjacent areas r _i , r _j is:

The color histogram of the new region r _ij obtained by merging r _i and r _j is:

size(r _ij )=size(r _i )+size(r _j ) (33)

Where size(r _i ) is the size of the initial region r _i ;

(b) The calculation method of texture similarity is:

For each initial region, calculate its SIFT (scale-invariant feature transform) feature. In the present invention, the scale factor takes a value of 1 in this patent, and the histogram interval width is 10, obtaining obtained as texture features.

in, Indicates the k-th texture feature in the i-th initial region;

The texture similarity of two adjacent regions r _i and r _j is:

The texture histogram of the new region r _ij obtained by merging r _i and r _j is

(C) The calculation method of size similarity is:

Where size(im) represents the size of the entire grayscale image.

(d) The calculation method of matching similarity is:

where BB _ij is a bounding box that only encloses regions r _i and r _j .

Finally, the four similarities are combined to obtain the final similarity:

s(r _i ,r _j )＝a ₁ Scolor(r _i ,r _j )+a ₂ Texture(r _i ,r _j )+a ₃ Ssize(r _i ,r _j )+a ₄ Sfill(r _i ,r _j )(38)

where a _i ∈ {0,1}, indicates whether the corresponding similarity is adopted.

(3.3) All the similarities form a similarity set S, and the two adjacent areas with the largest similarity s(r _i , r _j )=max(S) in the similarity set, merge these two areas r _ij =r _i ∪r _j , calculate the similarity between the new area r _ij and the adjacent area, update the similarity set: delete all the similarities s(r _i ,r _* ) related to the merged area r _i , and All similarities s(r _j , r _# ) related to the merged region r _j , added to the updated similarity set is: add the similarity between the new region r _ij and the adjacent region;

(3.4) Repeat steps 3.2-3.3, when the similarity set S is an empty set, the loop ends. Complete the merging of the initial regions to obtain the candidate region set R;

(3.5) Select the upper and lower chip regions to be analyzed from the set of candidate regions R, as shown in Fig. The regions to be analyzed need to contain respectively and For these 4 key geometric points, the resolution of the packaged original film is a*b.

(3.4) The loop ends and the candidate region R is obtained; the above selective search algorithm steps are performed on the grayscale image shown in Figure 4 to obtain 214 candidate regions, some of which are shown in Figure 5 .

(3.5) Region screening: filter conditions are set, and the region of the chip to be analyzed is selected from the candidate region R. Screening the grayscale image shown in Figure 4 will result in the area to be analyzed as shown in the images on the right side of Figure 6.

(4) Carry out threshold-based image segmentation to the filtered region to be analyzed, this step is realized by the following sub-steps:

(4) Carry out threshold-based image segmentation to the filtered region to be analyzed, this step is realized by the following sub-steps:

(4.1) Calculate the normalized histogram of the area to be analyzed. Let {0,1,...,L-1} represent L different gray levels in the area image with size M*N pixels, and n _z represent the number of pixels with gray level z. The total number of pixels in the region image is MN=n ₀ +n ₁ +n ₂ +...+n _L-1 , p _z =n _z /MN represents the normalized frequency of the gray level z in the histogram , z=0,1,2,...,L-1.

(4.2) For u=0,1,2,...,L-1, calculate the cumulative sum P ₁ (u), the calculation formula is as follows:

(4.3) For u=0,1,2,...,L-1, calculate the cumulative mean value m(u), the calculation formula is as follows:

(4.4) Calculating the global gray mean value m _G , the calculation formula is as follows:

(4.5) For u=0,1,2,...,L-1, calculate the variance between classes Calculated as follows:

(4.6) obtain the Otsu threshold u ^* by making the objective function (13) obtain the maximum value, if the maximum value is not unique, obtain u ^* with the average of each maximum value u detected correspondingly;

(4.7) Segment the region image based on the optimal threshold u ^* using the following formula to obtain the segmented image:

f(x, y) is the pixel of the pixel point (x, y) in the image of the area to be analyzed. After the segmentation, the pixel point of 1 represents the detected overflow part in the production process. The black part in the dotted line box in Figure 7 is the flashing component in the chip, and the flashing situation can be seen intuitively. If the flashing is serious, the packaging process should be stopped immediately, and the mold parameters should be checked to avoid more flashing.

It should be understood that the present invention, as a detection algorithm, is not limited to the specific experimental facilities and experimental conditions of the above-mentioned specific examples, and those skilled in the art can also make equivalent deformations or replacements without violating the spirit of the present invention. , these equivalent modifications or replacements are all included within the scope defined by the claims of the present application.

Claims (5)

1. A semiconductor chip flash rapid detection method based on an image segmentation algorithm is characterized by comprising the following steps:

(1) acquiring a multi-frame gray image of a packaging original sheet by using an industrial camera;

(2) performing Gaussian blur preprocessing on each frame of image;

(3) based on selective search image segmentation, hundreds of candidate regions are obtained and chip regions to be analyzed are screened out, and the method is realized by the following substeps:

(3.1) on a single frame imageOver-dividing to obtain n initial regions r₁,r₂,...,r_nForming an initial region set R ═ { R ═ R₁,r₂,...,r_n}; meanwhile, an initialization similarity set S is constructed,

(3.2) calculating the similarity s (r) of every two adjacent areas_i,r_j) The similarity index adopts color similarity, size similarity, texture similarity, coincidence similarity and four complementary similarities, and the four complementary similarities are respectively combined for the areas from different aspects;

(a) the color similarity is calculated as follows:

dividing each initial region into e intervals according to pixel values to obtain a normalized color histogramWherein,the normalization frequency number of the ith pixel value interval in the ith initial area is represented;

two adjacent regions r_i,r_jThe color similarity of (a) is:

merge r_i,r_jThe new region r obtained_ijThe color histogram is:

size(r_ij)＝size(r_i)+size(r_j) (4)

wherein size (r)_i) Is an initial region r_iThe size of (d);

(b) the texture similarity calculation method comprises the following steps:

calculating the SIFT (scale-invariant feature transform) feature of each initial region to obtainAnd obtaining the number of the texture features as v. Wherein,representing the ith texture feature in the ith initial region;

two adjacent regions r_i,r_jThe texture similarity of (a) is:

merge r_i,r_jThe new region r obtained_ijThe histogram of the texture is

(C) The calculation method of the size similarity comprises the following steps:

where size (im) represents the size of the entire grayscale image.

(d) The calculation method of the coincidence similarity comprises the following steps:

wherein BB_ijTo surround only the region r_iAnd r_jThe bounding box of (1).

And finally, integrating the four similarities to obtain the final similarity:

s(r_i,r_j)＝a₁Scolor(r_i,r_j)+a₂Stexture(r_i,r_j)+a₃Ssize(r_i,r_j)+a₄Sfill(r_i,r_j)(9)

wherein a is_iE {0,1}, indicating whether the corresponding similarity is adopted.

(3.3) all the similarities form a similarity set S, and the two adjacent regions S (r) with the highest similarity in the similarity set_i,r_j) Combine these two regions r max (S)_ij＝r_i∪r_jCalculating the new region r_ijAnd (3) updating the similarity set with the similarity of adjacent regions: deleting the region r with which the data is merged_iAll similarities s (r) of the correlation_i,r_*) And with the region r being merged_jAll similarities s (r) of the correlation_j,r_#) And adding the updated similarity set as follows: adding a new region r_ijSimilarity to adjacent regions;

and (3.4) circulating the step 3.2-3.3, and ending the circulation when the similarity set S is an empty set. Completing the combination of the initial regions to obtain a candidate region set R;

(3.5) screening an upper chip area and a lower chip area to be analyzed from the candidate area set R, so that the size of the area to be analyzed is in the range of one third to one half of the size of the original packaging piece, and the two areas to be analyzed need to be respectively containedAndthe resolution of the 4 key geometric points and the packaging original sheet is a x b.

(4) Performing threshold-based image segmentation on the screened region to be analyzed, wherein the step is realized by the following substeps:

(4.1) calculating a normalized histogram of the region to be analyzed. Let {0, 1.,. L-1} denote L different gray levels in a region image of size M N pixels, N_zIndicating the number of pixels with a gray level z. The total number of pixels in the area image is MN ═ n₀+n₁+n₂+...+n_L-1，p_z＝n_zthe/MN indicates a normalization frequency of z, which is a gray level in the histogram, and is 0,1, 2.

(4.2) for u-0, 1,2₁(u), the calculation formula is as follows:

(4.3) for u 0,1, 2.

(4.4) calculating the Global Gray mean m_GThe calculation formula is as follows:

(4.5) for u-0, 1,2The calculation formula is as follows:

(4.6) obtaining the Otsu threshold u by maximizing the objective function (13)^*If the maximum values are not unique, u is obtained by averaging the respective detected maximum values u^*；

(4.7) based on the optimal threshold u^*The region image is segmented by the following formula to obtain a segmented image:

f (x, y) is the pixel of the pixel point (x, y) in the image of the area to be analyzed. And the segmented pixels are 1 and represent the detected flash part in the production process.

2. The method according to claim 1, wherein in step 2.1, the larger σ, the wider the band of gaussian blur, the better the smoothness of the image and the higher the loss of detail. By adjusting the sigma parameter, the cancellation of image noise and preservation of integrity can be balanced.

3. The method of claim 1, wherein the packaging raw sheet in step 1 is a packaging raw sheet produced by using a semiconductor packaging press (FSAM120-1US) and a qfnqqnb 7X7-48L (T0.75) die.

4. The method of claim 1, wherein the gaussian blur step of step 2 is as follows:

(2.1) generating a gaussian operand of size (2k +1) × (2k +1) according to equation (1):

wherein k is an integer, w_pqIs the weight of the (p, q) coordinate position in the gaussian operand, and σ is the standard deviation of the gaussian distribution.

(2.2) calculating the reciprocal of the sum of Gaussian operand coefficientsAnd multiplying each element in the Gaussian operand by the element to obtain a new Gaussian operand.

And (2.3) reading pixels from the image and performing convolution operation on the pixels and the Gaussian operand of the second step, thereby completing the Gaussian blur preprocessing.

5. The method according to claim 1, wherein the gaussian blur of step 2 removes unnecessary noise while preserving the integrity of the image information, and reduces the resolution of the image to speed up the detection of the flash analysis.