CN103632343A - Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging - Google Patents
Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging Download PDFInfo
- Publication number
- CN103632343A CN103632343A CN201310561417.0A CN201310561417A CN103632343A CN 103632343 A CN103632343 A CN 103632343A CN 201310561417 A CN201310561417 A CN 201310561417A CN 103632343 A CN103632343 A CN 103632343A
- Authority
- CN
- China
- Prior art keywords
- image
- lnf
- small echo
- frequency field
- transformation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
The invention discloses a wavelet homomorphic filtering method for an X-ray image in integrated circuit packaging. The method comprises the following steps of (1) acquiring an X-ray image of a packaged component to obtain an X-ray image f(x, y); (2) performing logarithmic transformation on the X-ray image f(x, y) to obtain a transformed image lnf(x, y); (3) transforming the image lnf(x, y) from a space domain to a frequency domain in a wavelet transformation way to obtain a frequency-domain image F(s, t); (4) selecting a homomorphic filtering function H(s, t), and performing homomorphic filtering on the frequency-domain image to obtain a wavelet homomorphic filtered image G(s, t); (5) performing wavelet reverse transformation on the frequency-domain image G(s, t) to obtain a space-domain image g(x, y); (6) performing exponential transformation on the image g(x, y) to obtain a final processed image ff(x, y). According to the method, noise is effectively removed from the X-ray image in an integrated circuit packaging process; moreover, the method has the advantages of high calculation speed and convenience in operation.
Description
Technical field
The present invention relates to image and process research field, particularly a kind of radioscopic image small echo homographic filtering method for integrated antenna package.
Background technology
IC industry has become one of mainstay of national economy and national defense construction, is the important symbol of weighing a country scientific and technological strength.The integrated circuit especially production that is fabricated onto finished product of terminal company of great scale integrated circuit has experienced series of complex operation.Along with the development of IC (Integrated Circuit Design) technology, for the inside of components and parts, detect and propose higher requirement.
In encapsulation process, pad long term exposure is in air, easily produce oxidation, thereby all may there is defect in all tie points: comprise connection welding crack, there is no to connect that cavity upper, solder joint is too much, the defect of wire and wire pressure welding, and the problem of nude film and linkage interface.In addition, pad silicon chip can be because pressure produces fine crack in encapsulation process, and the colloid that conducting resinl connects also can produce bubble in encapsulation process, and the existence of bubble will have a strong impact on mechanical property, the electrical specification of components and parts.These all can exert an influence to the package quality of great scale integrated circuit.And the surperficial invisible defects such as salient point connection, tie point rosin joint and silicon chip fine crack, colloid bubble all cannot judge by AOI technology.
X ray, because of the transmission performance of itself, can detect surperficial invisible defect, so X ray detection obtains applying more and more widely.Yet, what in integrated antenna package testing process, obtain is dynamic image, be the result of X ray particle cumulative statistics on flat panel detector, imaging process is subject to the impact of the various factorss such as X ray scattering, electronic noise, light quantum noise, makes radioscopic image with much noise.Therefore, need first radioscopic image to be carried out to denoising, make the image after processing be more conducive to follow-up defects detection.
At present, conventional image filtering method is divided into spatial domain image filtering and frequency domain image filtering two classes.Spatial domain image filtering is divided into smooth Spaces wave filter and sharpening spatial filter; Frequency domain filtering method mainly comprises low-pass filter, bandpass filter, Hi-pass filter.But these filtering methods all can not carry out denoising to radioscopic image well, and can not meet the accuracy requirement of defects detection.Therefore,, in order to improve the radioscopic image defects detection precision of integrated antenna package process, just must invent new method, to improve detection speed and accuracy.
Summary of the invention
Fundamental purpose of the present invention is that the shortcoming that overcomes prior art is with not enough, a kind of radioscopic image small echo homographic filtering method for integrated antenna package is provided, the method can effectively be removed the noise of the radioscopic image in integrated antenna package, and have advantages of simple, quick, accuracy is high.
Object of the present invention realizes by following technical scheme: a kind of radioscopic image small echo homographic filtering method for integrated antenna package, comprises the following steps:
(1) gather the radioscopic image of packed components and parts, obtain radioscopic image f (x, y);
(2) described radioscopic image f (x, y) is carried out to log-transformation, obtain changing image lnf (x, y);
(3) by wavelet transformation, image lnf (x, y) is transformed from a spatial domain to frequency field, obtain frequency field image F (s, t);
(4) select homomorphic filtering function H (s, t), frequency field image is carried out to homomorphic filtering, obtain image G (s, t) after small echo homomorphic filtering;
(5) frequency field image G (s, t) is carried out to inverse wavelet transform, obtain spatial domain image g (x, y);
(6) image g (x, y) is carried out to exponential transform, obtain last processing image ff (x, y).
The method of preferably, carrying out log-transformation in described step (2) is:
(2-1) by radioscopic image f (x, y) model representation, be:
f(x，y)=i(x，y)*r(x，y)；
Wherein, i (x, y) represents luminance component, and r (x, y) represents reflectivity component;
(2-2) f (x, y) is carried out to log-transformation, it is become to additive component from the property taken advantage of component, obtain:
lnf(x，y)=lni(x，y)+lnr(x，y)。
Concrete, step (3) is described transforms from a spatial domain to frequency field by wavelet transformation by image lnf (x, y), obtains frequency field image F (s, t), and method is:
Adopt wavelet transformation:
WT{lnf(x，y)}=WT{lni(x，y)}+WT{lnr(x，y)}；
Make F (s, t)=WT{lnf (x, y) }, I (s, t)=WT{lni (x, y) }, R (s, t)=WT{lnr (x, y) }, above-mentioned formula brief note is:
F(s，t)=I(s，t)+R(s，t)；
In formula: I (s, t) represents frequency field illumination function; R (s, t) represents frequency field reflective function.
Concrete, step (5) is described carries out inverse wavelet transform to frequency field image G (s, t), obtains spatial domain image g (x, y), and method is:
g(x，y)=WT
-1(G(s，t))。
Concrete, step (6) is described carries out exponential transform to image g (x, y), obtains last processing figure ff (x, y), is specially:
ff(x，y)=exp(g(x，y))。
Preferably, homomorphic filtering function H (s, t) described in step (4) chooses, and is specially: H (s, the t) >1 of low frequency part, H (s, the t) <1 of HFS.
Preferably, described in step (4), frequency domain figure is looked like to carry out small echo homomorphic filtering, is specially:
G(s，t)=H(s，t)·F(s，t)=H(s，t)·I(s，t)+H(s，t)·R(s，t)。
Compared with prior art, the present invention has the following advantages and beneficial effect: the present invention is directed to signal to noise ratio (S/N ratio) low, radioscopic image in the poor integrated antenna package of defect contrast, the characteristic of picture noise itself and the feature of image itself have been considered, effectively removed the noise of the radioscopic image in integrated antenna package, and algorithm of the present invention is simple, quick, has higher accuracy.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the inventive method.
Embodiment
Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited to this.
As shown in Figure 1, for the radioscopic image small echo homographic filtering method of integrated antenna package, comprise the following steps described in the present embodiment:
(1) gather the radioscopic image of packed components and parts, obtain radioscopic image f (x, y).
(2) described radioscopic image f (x, y) is carried out to log-transformation, obtains changing image lnf (x, y), be specially:
(2-1) by radioscopic image f (x, y) model representation, be:
f(x，y)=i(x，y)*r(x，y)
Wherein, i (x, y) luminance component, r (x, y) reflectivity component.Due to illumination, variation is very little relatively, can be regarded as the low frequency composition of image, and reflectivity is high frequency composition.By processing respectively illumination and reflectivity component gray-scale value, reach the object of removing picture noise.
(2-2) f (x, y) is carried out to log-transformation, it is become to additive component from the property taken advantage of component:
lnf(x，y)=lni(x，y)+lnr(x，y)。
(3) by wavelet transformation, image lnf (x, y) is transformed from a spatial domain to frequency field, obtains frequency field image F (s, t), be specially:
Wavelet transformation:
WT{lnf(x，y)}=WT{lni(x，y)}+WT{lnr(x，y)}
Make F (s, t)=WT{lnf (x, y) }, I (s, t)=WT{lni (x, y), R (s, t)=WT{lnr (x, y) } brief note be:
F(s，t)=I(s，t)+R(s，t)
In formula: I (s, t)-illumination function spatially converts slowly, and its spectral characteristic concentrates on low-frequency range; The spectrum concentration of R (s, t)-reflective function is at high-frequency region.
(4) select homomorphic filtering function H (s, t), frequency field image carried out to homomorphic filtering, obtain image G (s, t) after small echo homomorphic filtering, comprise the following steps:
(4-1) choosing of homomorphic filtering function H (s, t):
Can image reach final filter effect and depend on choosing of homomorphic filtering function H (s, t), and the present embodiment designs one to high and low frequency component being affected to different homomorphic filtering function H (s, t).Because enhancing luminance component can expand the DYNAMIC DISTRIBUTION of gradation of image, diminished reflex component can be removed picture noise, so homomorphic filtering function H (s, t) selection is as follows: the H (s of low frequency part, t) >1, H (s, the t) <1 of HFS.
(4-2) frequency domain figure is looked like to carry out small echo homomorphic filtering, is specially:
G(s，t)=H(s，t)·F(s，t)=H(s，t)·I(s，t)+H(s，t)·R(s，t)。
(5) frequency field image G (s, t) is carried out to inverse wavelet transform, obtains spatial domain image g (x, y), be specially:
g(x，y)=WT
-1(G(s，t))。
(6) image g (x, y) is carried out to exponential transform, obtains last processing figure ff (x, y), be specially:
ff(x，y)=exp(g(x，y))。
Above-described embodiment is preferably embodiment of the present invention; but embodiments of the present invention are not limited by the examples; other any do not deviate from change, the modification done under Spirit Essence of the present invention and principle, substitutes, combination, simplify; all should be equivalent substitute mode, within being included in protection scope of the present invention.
Claims (7)
1. for a radioscopic image small echo homographic filtering method for integrated antenna package, comprise the following steps:
(1) gather the radioscopic image of packed components and parts, obtain radioscopic image f (x, y);
(2) described radioscopic image f (x, y) is carried out to log-transformation, obtain changing image lnf (x, y);
(3) by wavelet transformation, image lnf (x, y) is transformed from a spatial domain to frequency field, obtain frequency field image F (s, t);
(4) select homomorphic filtering function H (s, t), frequency field image is carried out to homomorphic filtering, obtain image G (s, t) after small echo homomorphic filtering;
(5) frequency field image G (s, t) is carried out to inverse wavelet transform, obtain spatial domain image g (x, y);
(6) image g (x, y) is carried out to exponential transform, obtain last processing image ff (x, y).
2. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 1, is characterized in that, the method for carrying out log-transformation in described step (2) is:
(2-1) by radioscopic image f (x, y) model representation, be:
f(x，y)=i(x，y)*r(x，y)；
Wherein, i (x, y) represents luminance component, and r (x, y) represents reflectivity component;
(2-2) f (x, y) is carried out to log-transformation, it is become to additive component from the property taken advantage of component, obtain:
lnf(x，y)=lni(x，y)+lnr(x，y)。
3. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 2, it is characterized in that, step (3) described by wavelet transformation by image lnf (x, y) transform from a spatial domain to frequency field, obtain frequency field image F (s, t), method is:
Adopt wavelet transformation:
WT{lnf(x，y)}=WT{lni(x，y)}+WT{lnr(x，y)}；
Make F (s, t)=WT{lnf (x, y) }, I (s, t)=WT{lni (x, y) }, R (s, t)=WT{lnr (x, y) }, above-mentioned formula brief note is:
F(s，t)=I(s，t)+R(s，t)；
In formula: I (s, t) represents frequency field illumination function; R (s, t) represents frequency field reflective function.
4. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 3, is characterized in that, step (5) is described to frequency field image G (s, t) carry out inverse wavelet transform, obtain spatial domain image g (x, y), method is:
g(x，y)=WT
-1(G(s，t))。
5. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 1, is characterized in that, step (6) is described to image g (x, y) carry out exponential transform, obtain last processing figure ff (x, y), be specially:
ff(x，y)=exp(g(x，y))。
6. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 1, it is characterized in that, (the s of homomorphic filtering function H described in step (4), choosing t), be specially: the H (s of low frequency part, t) >1, H (s, the t) <1 of HFS.
7. the radioscopic image small echo homographic filtering method for integrated antenna package according to claim 3, is characterized in that, described in step (4), frequency domain figure is looked like to carry out small echo homomorphic filtering, is specially:
G(s，t)=H(s，t)·F(s，t)=H(s，t)·I(s，t)+H(s，t)·R(s，t)。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310561417.0A CN103632343A (en) | 2013-11-12 | 2013-11-12 | Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310561417.0A CN103632343A (en) | 2013-11-12 | 2013-11-12 | Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103632343A true CN103632343A (en) | 2014-03-12 |
Family
ID=50213361
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310561417.0A Pending CN103632343A (en) | 2013-11-12 | 2013-11-12 | Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103632343A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006113813A2 (en) * | 2005-04-15 | 2006-10-26 | Invivo Corporation | Method for image intensity correction using extrapolation and adaptive smoothing |
| CN102567743A (en) * | 2011-12-20 | 2012-07-11 | 东南大学 | Automatic identification method of driver gestures based on video images |
-
2013
- 2013-11-12 CN CN201310561417.0A patent/CN103632343A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006113813A2 (en) * | 2005-04-15 | 2006-10-26 | Invivo Corporation | Method for image intensity correction using extrapolation and adaptive smoothing |
| CN102567743A (en) * | 2011-12-20 | 2012-07-11 | 东南大学 | Automatic identification method of driver gestures based on video images |
Non-Patent Citations (2)
| Title |
|---|
| 吕绪良 等: "基于小波变换的红外图像模糊与同态增强", 《解放军理工大学学报（自然科学版）》 * |
| 李红云 主编: "《数字图像处理》", 31 May 2012, 北京大学出版社 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103500442A (en) | X-ray image multi-scale detail enhancement method in integrated circuit packaging | |
| Wang et al. | An experiment-based review of low-light image enhancement methods | |
| JP4766118B2 (en) | Image processing method | |
| CN108090886B (en) | High dynamic range infrared image display and detail enhancement method | |
| JP2019536128A5 (en) | ||
| JP2011521387A (en) | Isolated edge enhancement architecture | |
| CN103489223A (en) | 3D tomoscan imaging method for 3D packaged IC | |
| Joseph et al. | Mammogram image denoising filters: A comparative study | |
| Mustafa et al. | Illumination correction of retinal images using superimpose low pass and Gaussian filtering | |
| CN103632342A (en) | Fuzzy enhancement method for X-ray image in integrated circuit packaging | |
| JP2014225787A (en) | Noise reduction processor | |
| CN109919922B (en) | Quality detection method for silicon-silicon direct bonding based on combination of spatial domain and morphology | |
| CN103632343A (en) | Wavelet homomorphic filtering method for X-ray image in integrated circuit packaging | |
| CN109724693A (en) | A kind of fusion spectrum denoising method based on stationary wavelet | |
| Makandar et al. | Comparative study of different noise models and effective filtering techniques | |
| Thriveni | Satellite image enhancement using discrete wavelet transform and threshold decomposition driven morphological filter | |
| CN113744163B (en) | Integrated circuit image enhancement method and device, electronic equipment and storage medium | |
| KR102190286B1 (en) | Image Processing Method for Enhancing Defects in Electronic Device Radioactive Inspection Apparatus | |
| CN111161162B (en) | Processing method and device for infrared image detail layer | |
| Artmann | Measurement of Noise using the dead leaves pattern | |
| Sakai et al. | Image improvement using image processing for scanning acoustic tomograph images | |
| Sun et al. | Adaptive bilateral filter considering local characteristics | |
| CN111855672A (en) | Method for detecting COF flexible board defects | |
| Kaur et al. | A novel approach for despeckling of ultrasound images | |
| Nnolim | Implementation of spatial domain Homomorphic filtering on embedded mobile devices |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140312 |