CN109461136A - The detection method of fiber distribution situation in a kind of blended fibre products - Google Patents
The detection method of fiber distribution situation in a kind of blended fibre products Download PDFInfo
- Publication number
- CN109461136A CN109461136A CN201811099487.8A CN201811099487A CN109461136A CN 109461136 A CN109461136 A CN 109461136A CN 201811099487 A CN201811099487 A CN 201811099487A CN 109461136 A CN109461136 A CN 109461136A
- Authority
- CN
- China
- Prior art keywords
- fiber
- image
- target fibers
- fibre products
- distribution situation
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N2021/8444—Fibrous material
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Image Processing (AREA)
- Nonwoven Fabrics (AREA)
Abstract
The invention discloses a kind of detection methods of fiber distribution situation in blended fibre products: one, acquiring blended fibre products original image, obtain low noise grayscale image through denoising and gray processing；Two, any fiber is set as target or non-targeted fiber；Three, " one " described image is cut to statuette sketch map, appropriate threshold is selected to handle it；Four, extracting non-targeted fiber characteristic value is noise, and " three " described image is done two-dimensional multistage accidental resonance, extracts target fibers according to result；Five, " four " described image is done into binaryzation, small figure is reassembled into life size binary picture by former cutting sequence, carried out if this binary picture can show target fibers profile in next step, otherwise return to " four "；Six, image object fiber pixel number after statistics is restored, provides blended fibre products fiber distribution situation numerical value.The present invention highlights target fibers by removing non-targeted fiber by extracting target and non-targeted fiber characteristic value, easy to operate, processing accuracy is high.
Description
Technical field
The present invention relates to Nonwoven Equipment field, it is especially applied to the inspection of fiber distribution situation in blended fibre products
It surveys.
Background technique
Sampling traditional fiber network detecting methods such as weight method, thickness measurement method and radioisotope method can energy
The uniformity of the accurate web for analyzing single fiber composition, but the web of composite fibre composition cannot be evaluated
Uniformity especially cannot intuitively embody the distribution situation of every kind of fiber in composite fibre net with image processing techniques
Continuous development, the processing accuracy having due to it is high, and reproducibility is good, and superiority, the technology such as processing diversification are used extensively
Fiber net structure feature is characterized, the main fiber orientation distribution research including non-woven web materials, web porosity
Multinomial application including detection, the measurement of web uniformity, has obtained preferable application effect still, above-mentioned image procossing
Method is more the performance indicator detection for the web of single fiber composition, and in the fiber of detection composite fibre composition
Fiber distribution situation in net then studies less this patent and combines multistage accidental resonance technology, passes through the method energy of image procossing
It effectively solves the problem above-mentioned, and has many advantages, such as rapidly and efficiently high with accuracy.
Currently, research fiber distribution characteristics method be usually used image co-registration to seek fiber orientation, in the form of
Corrosion, which calculates porosity corrosion with the miscellaneous point of expansion removal noise, with Threshold segmentation, so that profile and border is shunk；Expansion can
Make profile and border expansion burn into expansion be the gap in grid is removed to protrude fiber mesh, carrying out image threshold segmentation is to adopting
The one more appropriate threshold value of whole image selection collected is handled, and then the distribution situation of research fiber is still, also not
Have a kind of image processing techniques to study the fiber distribution situation in blended fibre products therefore, herein using it is multistage it is random altogether
The characteristic value of any one fiber in blended fibre products is set target fibers by the method for vibration, remaining is non-targeted fiber；
Then using image processing techniques come the distribution situation of qualitative and quantitative analysis target fibers；Finally according to point of target fibers
Cloth situation draws target fibers pixel distribution histogram.
Summary of the invention
Technical problem to be solved by the invention is to provide a kind of detection sides of fiber distribution situation in blended fibre products
Method can obtain the distribution situation of fiber in composite fibre net, with for qualitative and quantitative analysis composite fibre net structure and
Performance.
In order to solve the above technical problems, fiber is distributed feelings in a kind of blended fibre products of the technical solution adopted in the present invention
The detection method of condition, basic procedure as shown in Figure 1, specifically includes the following steps:
One, the original image for the product being made of multiple fiber mixing is acquired, and image is carried out at denoising and gray processing
Reason, obtains low noise grayscale image；
Two, can will be appointed according to actual demands of engineering using the photosensitive difference of every kind of fiber in above-mentioned blended fibre products
One or more fibers of anticipating are set as target fibers or non-targeted fiber.
Three, according to the actual accuracy requirement of different engineerings, aforementioned low noise grayscale image is integrally subjected to a certain size and ratio
Cutting, then according to the suitable threshold value of optimal imaging effect selection to each width cutting after image be respectively processed.
Four, characteristic value therein is extracted as noise, by treated each width to the non-targeted fiber set
Image input two-dimensional multistage connection accidental resonance after cutting is handled, and according to treated, result extracts target fibers；
Five, the image after said extracted to be gone out to every width cutting of target fibers carries out binary conversion treatment, then by all figures
As being reassembled into the binary picture of life size by former cutting sequence, if binary picture at this time can show target fibers profile after
It is continuous to carry out in next step, the 4th step, which is directly returned, if it cannot show target fibers profile re-starts processing；
Six, pixel number shared by image object fiber after statistics is restored, and provide target fibers in blended fibre products
With the index value of non-targeted fiber distribution situation.
Detailed description of the invention
Fig. 1 is the invention patent flow chart.
Fig. 2 is that glass/polyester fiber composite fibre net image original image is acquired in the present invention.
Fig. 3 is glass of the present invention/polyester fiber composite fibre net primitive beginning image denoising figure.
Fig. 4 be the present invention by negate and depression of order processing after glass/polyester fiber composite fibre net image.
Fig. 5 is small pixel cutting drawing after intercommunication rank of the present invention.
Fig. 6 is that threshold value of the present invention is small pixel image after 0.04 processing.
Fig. 7 is that threshold value of the present invention is small pixel image after 0.08 processing.
Fig. 8 is binary image after threshold value 0.04 of the present invention processing.
Fig. 9 is binary image after threshold value 0.08 of the present invention processing.
Figure 10 is binaryzation restored map of the present invention.
Figure 11 is the pixel number that each column of the present invention contains glass fibre.
Figure 12 is the pixel number that the every row of the present invention contains glass fibre.
Figure 13 is that basalt/polyester fiber composite fibre net image original image is acquired in the present invention.
Figure 14 is basalt of the present invention/polyester fiber composite fibre net original image denoising figure.
Figure 15 be the present invention by negate and depression of order processing after basalt/polyester fiber composite fibre net image.
Figure 16 is small pixel cutting drawing after intercommunication rank of the present invention.
Figure 17 is that threshold value of the present invention is small pixel image after 0.10 processing.
Figure 18 is that threshold value of the present invention is small pixel image after 0.06 processing.
Figure 19 is binary image after threshold value 0.10 of the present invention processing.
Figure 20 is binary image after threshold value 0.06 of the present invention processing.
Figure 21 is that accidental resonance of the present invention handles binaryzation restored map.
Figure 22 is that the secondary accidental resonance of the present invention handles binaryzation restored map.
Figure 23 is the pixel number that each column of the present invention contains basalt fibre.
Figure 24 is the pixel number that the every row of the present invention contains basalt fibre.
Specific embodiment
Embodiment 1:
By taking glass fibre/polyester fiber composite fibre net product as an example, fibre furnish and grouping are as shown in table 1, to say
The feasibility of bright this method, to illustrate the feasibility of this method.
Table 1
Firstly, with image acquisition device random acquisition fiber net image made of the 4th group of composite fibre proportion in table 1, as
Plain size is 3000 pixels × 4000 pixels (as shown in Figure 2), using Sobel filter method and weighted average method to collecting
Image carry out denoising and gray processing processing (as shown in Figure 3).
Since glass fibre and polyester fiber photoperceptivity are had any different, using the characteristic, white portion is glass fibers in Fig. 2
Dimension, black portions set target fibers for glass fibre herein according to actual demands of engineering for polyester fiber, polyester fiber is set
It is set to non-targeted fiber.
Collected original image is carried out first to negate depression of order processing using Matlab, as a result as shown in Figure 4；To depression of order
It is 32 × 52 small pixel images that image, which is cut into pixel size, afterwards, then as shown in Figure 5 to the greatest extent may be used on the image according to target fibers
Energy is clear and fidelity is strong, and non-targeted fiber shows principle as few as possible, and different thresholds are respectively adopted to small pixel image
Value is handled, as shown in Figure 6, Figure 7；Then to treated under different threshold values, image carries out binary conversion treatment, such as Fig. 8, Fig. 9 institute
Show.
The grayscale information characteristic value for extracting the non-targeted fiber set, is normalized place after noise is added to it again
Then reason respectively carries out it going being unfolded to be unfolded with column, and successively (parameter is respectively as follows: Gauss white noise for progress accidental resonance processing
Sound, noise intensity 4, h=0.1, a=6, b=12), it obtains two dimension and restores small image；
Small image obtained by the above method will be taken to carry out binary conversion treatment, all images are then pressed into former cutting sequence weight
Combination nova is at the binary picture of life size, and as shown in Figure 10, can be seen that from this figure can show target fibers profile, therefore not
It needs to return previous step and re-starts calculating；
Target fibers are every in binary image after image after binaryzation only has two kinds of black pixel and white pixel point, statistics to restore
The pixel number of row each column, the i.e. number of white pixel point draw histogram, as shown in Figure 11, Figure 12.
Embodiment two
It is also fine with basalt fibre/terylene herein to prove that this patent can be applied to other type blended fibre products
For tieing up composite fibre net product, fibre furnish and grouping are as shown in table 2.
Table 2
Image processing flow is as shown in Figure 1, the specific steps are as follows:
Firstly, with image acquisition device random acquisition fiber net image made of the 4th group of composite fibre proportion in table 2, as
Plain size is 3000 pixels × 4000 pixels (as shown in figure 13), using Sobel filter method and weighted average method to collecting
Image carry out denoising and gray processing processing (as shown in figure 14).
Since glass fibre and polyester fiber photoperceptivity are had any different, using the characteristic, white portion is the Black Warrior in Figure 13
Rock fiber, black portions set target fibers, terylene for basalt fibre herein according to actual demands of engineering for polyester fiber
Fiber is set as non-targeted fiber.
Collected original image is carried out first to negate depression of order processing using Matlab, as a result as shown in figure 15；To drop
After rank image be cut into pixel size be 32 × 52 small pixel images, as shown in figure 16 then according to target fibers on the image
As clear as possible and fidelity is strong, and non-targeted fiber shows principle as few as possible, and small pixel image is respectively adopted not
It is handled with threshold value, as shown in Figure 17, Figure 18；Then to treated under different threshold values, image carries out binary conversion treatment, such as schemes
19, shown in Figure 20.
The grayscale information characteristic value for extracting the non-targeted fiber set, is normalized place after noise is added to it again
Then reason respectively carries out it going being unfolded to be unfolded with column, and successively (parameter is respectively as follows: Gauss white noise for progress accidental resonance processing
Sound, noise intensity 4, h=0.1, a=6, b=12), it obtains two dimension and restores small image；Small image obtained by the above method will be taken
Binary conversion treatment is carried out, then all images are reassembled into the binary picture of life size by former cutting sequence, such as Figure 21 institute
Show, can be seen that can to show target fibers profile still unclear from this figure, it is therefore desirable to return to previous step and re-start calculating；
By recalculating, extract the grayscale information characteristic value of the non-targeted fiber set again, and it is added after noise again into
Then row normalized carries out row expansion to it respectively and is unfolded with column, and successively carry out accidental resonance processing (parameter is respectively
White Gaussian noise, noise intensity 4, h=0.1, a=6, b=12, second level accidental resonance parameter are h=0.1, a=5, b=
10) it, obtains two dimension and restores small image；Small image obtained by the above method will be taken to carry out binary conversion treatment, then by all figures
Binary picture as being reassembled into life size by former cutting sequence, as shown in figure 22, can be seen that from this figure can show mesh
Mark fiber profile, it is no longer necessary to return to previous step and re-start calculating.
Target is fine in binary image after image after binaryzation only has two kinds of black pixel and white pixel point, statistics to restore
The pixel number of each row and column, the i.e. number of white pixel point are tieed up, histogram is drawn, as shown in Figure 23, Figure 24.
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (3)
1. the detection method of fiber distribution situation in a kind of blended fibre products, which comprises the following steps:
One, the original image for the product being made of multiple fiber mixing is acquired, and denoising and gray processing processing are carried out to image, is obtained
To low noise grayscale image；
It two,, can will be any one using the photosensitive difference of every kind of fiber in above-mentioned blended fibre products according to actual demands of engineering
Kind or multiple fiber are set as target fibers or non-targeted fiber；
Three, according to the actual accuracy requirement of different engineerings, aforementioned low noise grayscale image is integrally subjected to cutting for a certain size and ratio
It cuts, then the image after the cutting of each width is respectively processed according to the suitable threshold value of optimal imaging effect selection；
Four, characteristic value therein is extracted as noise to the non-targeted fiber set, treated each width is cut
Image input two-dimensional multistage connection accidental resonance afterwards is handled, and according to treated, result extracts target fibers；
Five, the image after said extracted to be gone out to every width cutting of target fibers carries out binary conversion treatment, then presses all images
Former cutting sequence is reassembled into the binary picture of life size, continue if binary picture at this time can show target fibers profile into
Row in next step, directly returns to the 4th step if it cannot show target fibers profile and re-starts processing；
Six, pixel number shared by image object fiber after statistics is restored, and provide in blended fibre products target fibers with it is non-
The index value of target fibers distribution situation.
2. the detection method of fiber distribution situation in a kind of blended fibre products according to claim 1, it is characterised in that
Every piece image after original image cutting corresponds to an individual threshold value.
3. the detection method of fiber distribution situation in a kind of blended fibre products according to claim 1, it is characterised in that
Two-dimensional multistage connection stochastic resonance method can enhance its algorithm effect by using series is increased.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811099487.8A CN109461136B (en) | 2018-09-20 | 2018-09-20 | Method for detecting fiber distribution condition in mixed fiber product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811099487.8A CN109461136B (en) | 2018-09-20 | 2018-09-20 | Method for detecting fiber distribution condition in mixed fiber product |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109461136A true CN109461136A (en) | 2019-03-12 |
| CN109461136B CN109461136B (en) | 2022-09-27 |
Family
ID=65606837
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811099487.8A Active CN109461136B (en) | 2018-09-20 | 2018-09-20 | Method for detecting fiber distribution condition in mixed fiber product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109461136B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112164085A (en) * | 2020-09-28 | 2021-01-01 | 华南理工大学 | Fiber image segmentation and diameter statistical method based on image processing |
| CN113592811A (en) * | 2021-07-29 | 2021-11-02 | 常州大学 | Melt-blown fabric thickness consistency detection method based on image processing |
| CN113610852A (en) * | 2021-10-10 | 2021-11-05 | 江苏祥顺布业有限公司 | Yarn drafting quality monitoring method based on image processing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101241082A (en) * | 2008-03-14 | 2008-08-13 | 东华大学 | Nonwoven fibric fibre orientation distribution measuring systems and method |
| CN101320476A (en) * | 2008-07-04 | 2008-12-10 | 中国农业大学 | Cotton foreign fiber image processing system and method |
| CN102622602A (en) * | 2012-02-28 | 2012-08-01 | 中国农业大学 | Cotton foreign fiber image online dividing method and cotton foreign fiber image online dividing system |
| CN103020609A (en) * | 2012-12-30 | 2013-04-03 | 上海师范大学 | Complex fiber image recognition method |
| CN107343115A (en) * | 2016-05-02 | 2017-11-10 | 佳能株式会社 | Image processing apparatus, image processing method and non-transitory computer-readable storage media |
-
2018
- 2018-09-20 CN CN201811099487.8A patent/CN109461136B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101241082A (en) * | 2008-03-14 | 2008-08-13 | 东华大学 | Nonwoven fibric fibre orientation distribution measuring systems and method |
| CN101320476A (en) * | 2008-07-04 | 2008-12-10 | 中国农业大学 | Cotton foreign fiber image processing system and method |
| CN102622602A (en) * | 2012-02-28 | 2012-08-01 | 中国农业大学 | Cotton foreign fiber image online dividing method and cotton foreign fiber image online dividing system |
| CN103020609A (en) * | 2012-12-30 | 2013-04-03 | 上海师范大学 | Complex fiber image recognition method |
| CN107343115A (en) * | 2016-05-02 | 2017-11-10 | 佳能株式会社 | Image processing apparatus, image processing method and non-transitory computer-readable storage media |
Non-Patent Citations (8)
| Title |
|---|
| CYNIAK，D，等: ""Influence of selected parameters of the spinning process on the state of mixing of fibers of a cotton/polyester-fibre blend yarn"", 《FIBERS & TEXTTILES IN EASTERN EUROPE》 * |
| ZHANG HENG，等: ""Prediction of the Pore Size Distribution in Spunbond-Needlepunched Nonwoven Geotextiles Based on a Stochastic Simulation"", 《 2015 8TH INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTATION TECHNOLOGY AND AUTOMATION (ICICTA)》 * |
| 万永菁等: "一种用于纤维增强复合材料微观结构图像的阈值分割算法", 《华东理工大学学报(自然科学版)》 * |
| 冷永刚等: "二维随机共振参数调节的图像处理", 《天津大学学报》 * |
| 吴凯辉: ""水刺非织造Delta高速杂乱梳理成网工艺的研究"", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》 * |
| 彭伟良等: "山羊绒和细绵羊毛纤维的形态特征分析", 《上海纺织科技》 * |
| 杨潇等: "图像识别在羊绒羊毛检测中的应用研究", 《广西纺织科技》 * |
| 赵尔华: ""图像二维随机共振研究"", 《《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112164085A (en) * | 2020-09-28 | 2021-01-01 | 华南理工大学 | Fiber image segmentation and diameter statistical method based on image processing |
| CN112164085B (en) * | 2020-09-28 | 2023-08-18 | 华南理工大学 | Fiber image segmentation and diameter statistics method based on image processing |
| CN113592811A (en) * | 2021-07-29 | 2021-11-02 | 常州大学 | Melt-blown fabric thickness consistency detection method based on image processing |
| CN113592811B (en) * | 2021-07-29 | 2023-08-22 | 常州大学 | Melt-blown cloth thickness consistency detection method based on image processing |
| CN113610852A (en) * | 2021-10-10 | 2021-11-05 | 江苏祥顺布业有限公司 | Yarn drafting quality monitoring method based on image processing |
| CN113610852B (en) * | 2021-10-10 | 2021-12-10 | 江苏祥顺布业有限公司 | Yarn drafting quality monitoring method based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109461136B (en) | 2022-09-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10510148B2 (en) | Systems and methods for block based edgel detection with false edge elimination | |
| CN109461136A (en) | The detection method of fiber distribution situation in a kind of blended fibre products | |
| Carlile et al. | Determination of ecological scale | |
| CN107229918A (en) | A kind of SAR image object detection method based on full convolutional neural networks | |
| CN101441766B (en) | SAR image fusion method based on multiple-dimension geometric analysis | |
| CN101587189B (en) | Texture elementary feature extraction method for synthetizing aperture radar images | |
| CN112926429B (en) | Machine-check model training and video machine-check method, device, equipment and storage medium | |
| CN110533113A (en) | Branch's point detecting method of tree in a kind of digital picture | |
| CN111027343A (en) | Bar code area positioning method and device | |
| CN103778616A (en) | Contrast pyramid image fusion method based on area | |
| CN110176020A (en) | A kind of bird's nest impurity method for sorting merging 2D and 3D rendering | |
| Goyal et al. | An improved local binary pattern based edge detection algorithm for noisy images | |
| WO2019181072A1 (en) | Image processing method, computer program, and recording medium | |
| CN107871320A (en) | A kind of detecting system and method for insulating board heat ageing degree | |
| CN105608674B (en) | A kind of image enchancing method based on image registration, interpolation and denoising | |
| CN105205485B (en) | Large scale image partitioning algorithm based on maximum variance algorithm between multiclass class | |
| CN103455986B (en) | Random noise point detecting method based on fractional order differential gradient | |
| CN103778615B (en) | Multi-focus image fusing method based on region similitude | |
| CN108198192A (en) | A kind of quick human body segmentation's method of high-precision based on deep learning | |
| CN110378186B (en) | SAR remote sensing image ship target false alarm rejection method | |
| Strand | Detection of observer bias in ongoing forest health monitoring programmes | |
| CN112613521B (en) | Multilevel data analysis system and method based on data conversion | |
| CN115393796A (en) | X-ray security inspection image dangerous article detection method based on attention mechanism | |
| Hu et al. | Infrared and visible image fusion based on multiscale decomposition with Gaussian and co-occurrence filters | |
| Liu et al. | Adaptive ideal image reconstruction for bacteria colony detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |