CN104156951B - A kind of white blood cell detection method for BAL fluid smear - Google Patents
A kind of white blood cell detection method for BAL fluid smear Download PDFInfo
- Publication number
- CN104156951B CN104156951B CN201410369635.9A CN201410369635A CN104156951B CN 104156951 B CN104156951 B CN 104156951B CN 201410369635 A CN201410369635 A CN 201410369635A CN 104156951 B CN104156951 B CN 104156951B
- Authority
- CN
- China
- Prior art keywords
- region
- gray
- binary image
- image
- connected region
- 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.)
- Active
Links
Abstract
A kind of white blood cell detection method for bronchoalveolar lavage smear of the invention, the automatic testing method of leucocyte in a kind of smear for bronchoalveolar lavage, leucocyte automatic testing method in the bronchoalveolar lavage smear more particularly to based on digital image processing techniques.Gray processing, binary conversion treatment are carried out by the micro-image that bronchoalveolar lavage smear is gathered to microscope, screened with internal feature using the resemblance of leucocyte simultaneously, finally identify leucocyte, thus with easy to operate, detection efficiency is high, high precision, loss and false drop rate is low, low cost effect.
Description
Technical field
The present invention relates to a kind of automatic testing method of leucocyte in smear for BAL fluid, more particularly to
Leucocyte automatic testing method in BAL fluid smear based on digital image processing techniques.
Background technology
Digital picture refers to the digitized image obtained by digital imaging apparatus, and digital picture has high-resolution
The characteristics of with high gray value.And Digital Image Processing be by computer image is taken out noise, enhancing, recovery, segmentation,
The treatment technologies such as feature are extracted, to obtain human eye vision or certain image processing process required for receiving system.
The automatic detection of leucocyte is the position in microscope using digital image processing techniques analysis, positioning leucocyte, and is united
Count out the quantative attribute of leucocyte in field of microscope.
Leucocyte is divided into granulocyte, lymphocyte, monocyte, eosinophil, the type of basophilic granulocyte five.
At this stage, it is general using average gradient value is calculated for the leucocyte in blood, then carry out edge extracting and Threshold segmentation
Method detects that to the leucocyte in blood the method is often higher to the requirement of picture shooting condition, and bronchovesicular
Irrigating solution smear impurity is relatively more, and background condition is complicated, it is easy to influenceed detection to imitate by the interference of impurity or other cells
Really.
The content of the invention
The purpose of the present invention is directed to the deficiency of background technology, devises a kind of to the white of BAL fluid smear
Cell detection method, the method is based on digital image processing techniques, divides from the digital picture of BAL fluid smear
Leucocyte is discerned, so as to reach easy to operate, detection efficiency high, high precision, loss and false drop rate is low, low cost purpose.
The technical scheme that the present invention is provided is a kind of white blood cell detection method for BAL fluid smear, should
Method includes:
Step 1:The micro-image of BAL fluid smear is gathered using microscope;
Step 2:Gray processing treatment is carried out to the image in step 1, gray level image is converted into;
Step 3:Gray level image in step 2 is converted into binary image;
Step 4:Binary image in step 3 is carried out into denoising, edge enhancing and edge sharpening treatment;
Step 5:By the binary image connected component labeling after step 4 is processed, and calculate connected region and include:Face
Product, girth, eccentricity are in interior resemblance;
Step 6:Preliminary screening is carried out to the connected region in step 5 according to connected region resemblance, retains profile special
The connected region for meeting leucocyte condition is levied, is then divided into pyocyte cell suspicious region and common leucocyte according to its size
Suspicious region;
Different processing methods is used to this two classes cell suspicious region respectively, for common leucocyte suspicious region, I
Using step 7 arrive step 17 the step of:
Step 7:Common leucocyte suspicious region position in former gray-scale map in step 6 is found, and is cut in gray-scale map
Go out common leucocyte suspicious region, obtain some cutting images;
Step 8:Cutting image in step 7 is converted into binary image;
Step 9:Connected region is marked in the binary image that step 8 is obtained, count connected region area and
Centroid feature, finds the largest connected region of area, and exclude other connected regions away from the largest connected region barycenter of area;
Step 10:Image completion, expansion, the corrosion that step 9 is obtained, and connected region is re-flagged, count each connection
The area in region, finds largest connected region in each binary image, and calculate the circularity in largest connected region;
Step 11:Binary image is screened using largest connected zoned circular degree in each binary image, is retained
Meet the binary image of common leucocyte circularity feature;
Step 12:Each binary image that step 11 retains is filled, expansion process, re-flag connected region,
The area of connected region is counted, the largest connected region in each binary image is found, the circularity in largest connected region is calculated;
Step 13:It is largest connected in each binary image obtained to step 12 according to common leucocyte circularity feature
Zoned circular degree feature is screened, and further retains the binary image for meeting common white cell circularity feature;
Step 14:The feature of second largest connected region in each binary image that statistic procedure 13 is remained, calculates the
The circularity of two big connected regions and shared binary image area percentage;
Step 15:Circularity feature and shared binary image area hundred according to the second largest connected region of common leucocyte
Divide ratio, the circularity and area percentage in step 14 are screened to binary image, retain qualified binaryzation
Image, finds the binary image corresponding region, and mark in former gray-scale map;
Step 16:Average gray, the gray variance value of marked region are calculated, is put down using gray scale according to common leucocyte circle
Average and gray variance are further screened to the marked region that step 15 is obtained;
Step 17:Region after screening is marked in step 1 obtains image, and assert that the region is leucocyte pair
The region answered;
For pyocyte cell suspicious region, we using step 18 arrive step 24 the step of:
Step 18:The edge of pyocyte cell suspicious region in step 6 is found with Sobel Operator;
Step 19:The concave point at the pyocyte cell suspicious region edge that step 18 is obtained is found, and marks concave point position;
Step 20:According to the position of each concave point in step 19, pyocyte cell suspicious region is divided into it is multiple unicellular, and
Statistics number of cells;
Step 21:The number of cells for obtaining is counted by step 20, connected region of the number of cells less than 3 is excluded,;
Step 22:The corresponding region of reservation region in finding step 21 in former gray-scale map, and calculate average gray and
Gray variance value;
Step 23:The region found in step 22 is screened using average gray and gray variance;
Step 24:Region after screening is marked in artwork, and assert that the region is the corresponding region of leucocyte.
Further, the step 3 is concretely comprised the following steps:
Step 3-1:Gray threshold using maximum variance between clusters to gray level image to be obtained during setting binary conversion treatment
The gray threshold for arriving;
Step 3-2:Each pixel gray value in gray level image is compared with gray threshold, if being more than gray scale threshold
Value, then the gray scale is set to 0, if being less than gray threshold, this gray scale is set to 255, obtains binary image;
Further, connected component labeling of the area more than 1050 is pyocyte cell suspicious region, area in the step 6
Connected component labeling more than 350 and less than 1050 is common leucocyte suspicious region.
Further, the binary image of the circularity more than 0.6 in largest connected region is retained in the step 11.
Further, the binary image of the circularity more than 0.48 in largest connected region is retained in the step 13.
Further, second largest connected region circularity is retained in the step 15 to exist with area percentage less than 0.57
Binary image between 0.31 to 1.
Further, in the step 16 retain average gray be 100 to 180, gray variance value be 800 to 2000 it
Between region.
Further, the step 19 is concretely comprised the following steps:
Step 19-1:The tangent slope at edge is calculated using edge coordinate, and calculates the tangent slope change of every bit
Value;
Step 19-2:Scope discontinuity is found, and assert these tangent slope catastrophe points for concave point, and recording pit position
Put.
Further, it is 80 to 190 that average gray is retained in the step 23, and gray variance is the area of 2500-6000
Domain.
A kind of automatic discriminating conduct of leucocyte in smear for BAL fluid is invented herein, by micro-
The micro-image of mirror collection BAL fluid smear carries out gray processing, binary conversion treatment, while using the outer of leucocyte
Shape feature is screened with internal feature, finally identifies leucocyte, thus have that easy to operate, detection efficiency is high, high precision,
Loss and false drop rate is low, low cost effect.
Brief description of the drawings
Fig. 1 is the automatic discriminating conduct flow chart of leucocyte in a kind of smear for BAL fluid of the present invention;
Fig. 2 is the automatic discriminating conduct micro-image of leucocyte in a kind of picture for bronchoalveolar lavage of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawings, the automatic detection flow to leucocyte in a kind of BAL fluid smear of the invention is entered
Row is described in detail:
Step 1:The micro-image of BAL fluid smear is gathered using microscope;
Step 2:Gray processing treatment is carried out to the image in step 1, gray level image is converted into;
Step 3:Gray level image in step 2 is converted into binary image;
Step 3-1:Gray threshold using maximum variance between clusters to gray level image to be obtained during setting binary conversion treatment
The gray threshold for arriving;
Step 3-2:Each pixel gray value in gray level image is compared with gray threshold, if being more than gray scale threshold
Value, then the gray scale is set to 0, if being less than gray threshold, this gray scale is set to 255, obtains bianry image;
Step 4:Binary image in step 3 is carried out into denoising, edge enhancing and edge sharpening treatment;
Step 5:By the binary image connected component labeling in step 4, and calculate connected region and include:Area, girth, from
Heart rate is in interior resemblance;
Step 6:Preliminary screening is carried out to the UNICOM region in step 5 according to connected region resemblance, retains profile special
The connected region for meeting leucocyte condition is levied, is then divided into pyocyte cell suspicious region and common leucocyte according to its size
Suspicious region, the connected component labeling by area more than 1050 is pyocyte cell suspicious region, and area is more than 350 and less than 1050
Connected component labeling be common leucocyte suspicious region;
Different processing methods is used to this two classes cell suspicious region respectively, for common leucocyte suspicious region, I
Using step 7 arrive step 17 the step of:
Step 7:Common leucocyte suspicious region position in former gray-scale map in step 6 is found, and is cut in gray-scale map
Go out common leucocyte suspicious region, obtain some cutting images;
Step 8:Cutting image in step 7 is converted into binary image;
Step 9:Connected region is marked in the binary image that step 8 is obtained, count connected region area and
Centroid feature, finds the largest connected region of area, and exclude other connected regions away from the largest connected regional center of area;
Step 10:Image completion, expansion, the corrosion that step 9 is obtained, and connected region is re-flagged, count each connection
The area in region, finds largest connected region in each binary image, and calculate the circularity in largest connected region;
Step 11:Binary image is screened using largest connected zoned circular degree in each binary image, is retained
Binary image of the circularity in largest connected region more than 0.6;
Step 12:The binary image that step 11 retains is filled, expansion process, from new mark connected region, system
The area of connected region is counted, largest connected region is found, the circularity in largest connected region is calculated;
Step 13:The image in step 12 is screened using largest connected zoned circular degree feature, is further retained
Binary image of the circularity in largest connected region more than 0.48;
Step 14:The feature of second largest connected region in each binary image that statistic procedure 13 is remained, calculates the
The circularity of two big connected regions and shared binary image area percentage;
Step 15:Binary image is screened using the circularity and area percentage in step 14, retains second
The big binary image that connection zoned circular degree is less than 0.57 and area percentage is between 0.31 to 1;
Step 16:Average gray, the gray variance value of marked region are calculated, it is 100 to 180 to retain average gray,
Gray variance value is the region between 800 to 2000;
Step 17:Region after screening is marked in artwork, and assert that the region is the corresponding region of leucocyte;
For pyocyte cell suspicious region, we using step 18 arrive step 24 the step of:
Step 18:The edge of the pyocyte cell suspicious region marked in step 6 is found with Sobel Operator;
Step 19:The concave point at the pyocyte cell suspicious region edge that step 18 is obtained is found, and marks concave point position;
Step 19-1:The tangent slope at edge is calculated using edge coordinate, and calculates the tangent slope change of every bit
Value;
Step 19-2:Scope discontinuity is found, and assert these tangent slope catastrophe points for concave point, and recording pit position
Put;
Step 20:According to the position of each concave point in step 19, pyocyte cell suspicious region is divided into it is multiple unicellular, and
Statistics number of cells;
Step 21:The number of cells for obtaining is counted by step 20, connected region of the number of cells less than 3 is excluded;
Step 22:The corresponding region of reservation region in finding step 21 in former gray-scale map, and calculate average gray and
Gray variance value;
Step 23:The region found in step 22 is screened using average gray and gray variance, retains gray scale
Average value is 80 to 190, and gray variance is the region of 2500-6000;
Step 24:Region after screening is marked in artwork, and assert that the region is the corresponding region of leucocyte.
Claims (9)
1. a kind of white blood cell detection method for BAL fluid smear, the method includes:
Step 1:The micro-image of BAL fluid smear is gathered using microscope;
Step 2:Gray processing treatment is carried out to the micro-image in step 1, gray level image is converted into;
Step 3:Gray level image in step 2 is converted into binary image;
Step 4:Binary image in step 3 is carried out into denoising, edge enhancing and edge sharpening treatment;
Step 5:By the binary image connected component labeling after step 4 is processed, and calculate connected region and include:Area, week
Long, eccentricity is in interior resemblance;
Step 6:Preliminary screening is carried out to the connected region in step 5 according to connected region resemblance, retains resemblance symbol
The connected region of leucocyte condition is closed, is then divided into pyocyte cell suspicious region according to its size and common leucocyte is doubtful
Region;
Different processing methods is used to this two classes cell suspicious region respectively, for common leucocyte suspicious region, using step
Rapid 7 to step 17 the step of:
Step 7:Common leucocyte suspicious region position in gray level image in step 6 is found, and is cut out in gray level image
Common leucocyte suspicious region, obtains some cutting images;
Step 8:Cutting image in step 7 is converted into binary image;
Step 9:Connected region is marked in the binary image that step 8 is obtained, and counts the area and barycenter of connected region
Feature, finds the largest connected region of area, and exclude other connected regions away from the largest connected region barycenter of area;
Step 10:Image completion, expansion, the corrosion that step 9 is obtained, and connected region is re-flagged, count each connected region
Area, find largest connected region in each binary image, and calculate the circularity in largest connected region;
Step 11:Binary image is screened using largest connected zoned circular degree in each binary image of step 10,
Reservation meets the binary image of common white cell circularity feature;
Step 12:Each binary image that step 11 retains is filled, expansion process, re-flag connected region, count
The area of connected region, finds the largest connected region in each binary image, calculates the circularity in largest connected region;
Step 13:According to common leucocyte circularity feature, largest connected region in each binary image obtained to step 12
Circularity feature is screened, and further retains the binary image for meeting common white cell circularity feature;
Step 14:The feature of second largest connected region in each binary image that statistic procedure 13 is remained, calculates second largest
The circularity of connected region and shared binary image area percentage;
Step 15:Circularity feature and shared binary image area percentage according to the second largest connected region of common leucocyte
Than, the circularity and area percentage in step 14 are screened to binary image, retain qualified binary picture
Picture, finds the binary image corresponding region, and mark in gray level image;
Step 16:Average gray, the gray variance value of marked region are calculated, using average gray and gray variance to step
15 marked regions for obtaining further are screened;
Step 17:Region after screening is marked in step 1 obtains image, and assert the region for leucocyte is corresponding
Region;
For pyocyte cell suspicious region, the step of arrive step 24 using step 18:
Step 18:The edge of pyocyte cell suspicious region in step 6 is found with Sobel Operator;
Step 19:The concave point at the pyocyte cell suspicious region edge that step 18 is obtained is found, and marks concave point position;
Step 20:According to the position of each concave point in step 19, pyocyte cell suspicious region is divided into multiple unicellular, and counted
Number of cells;
Step 21:The number of cells for obtaining is counted by step 20, connected region of the number of cells less than 3 is excluded;
Step 22:The corresponding region of the connected region retained in finding step 21 in gray level image, and calculate average gray
With gray variance value;
Step 23:The region found in step 22 is screened using average gray and gray variance;
Step 24:Region after screening is marked in artwork, and assert that the region is the corresponding region of leucocyte.
2. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
In concretely comprising the following steps for step 3:
Step 3-1:Gray threshold is obtained to gray level image using maximum variance between clusters during setting binary conversion treatment
Gray threshold;
Step 3-2:Each pixel gray value in gray level image is compared with gray threshold, if being more than gray threshold,
The gray scale is set to 0, if being less than gray threshold, this gray scale is set to 255, obtains binary image.
3. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
Connected component labeling of the area more than 1050 is pyocyte cell suspicious region in step 6, and area is more than 350 and less than 1050
Connected component labeling is common leucocyte suspicious region.
4. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
Retain the binary image of the circularity more than 0.6 in largest connected region in step 11.
5. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
Retain the binary image of the circularity more than 0.48 in largest connected region in step 13.
6. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
Retain the binary picture that second largest connected region circularity is less than 0.57 and area percentage is between 0.31 to 1 in step 15
Picture.
7. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
It is 100 to 180 to retain average gray in step 16, and gray variance value is the region between 800 to 2000.
8. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
In concretely comprising the following steps for step 19:
Step 19-1:The tangent slope at edge is calculated using edge coordinate, and calculates the tangent slope changing value of every bit;
Step 19-2:Scope discontinuity is found, and assert these tangent slope catastrophe points for concave point, and recording pit position.
9. a kind of white blood cell detection method for BAL fluid smear as claimed in claim 1, its feature exists
It is 80 to 190 to retain average gray in step 23, and gray variance is the region of 2500-6000.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410369635.9A CN104156951B (en) | 2014-07-30 | 2014-07-30 | A kind of white blood cell detection method for BAL fluid smear |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410369635.9A CN104156951B (en) | 2014-07-30 | 2014-07-30 | A kind of white blood cell detection method for BAL fluid smear |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104156951A CN104156951A (en) | 2014-11-19 |
CN104156951B true CN104156951B (en) | 2017-06-16 |
Family
ID=51882441
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410369635.9A Active CN104156951B (en) | 2014-07-30 | 2014-07-30 | A kind of white blood cell detection method for BAL fluid smear |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104156951B (en) |
Families Citing this family (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104700112B (en) * | 2015-02-02 | 2018-07-27 | 电子科技大学 | Parasite egg detecting method in a kind of excrement based on morphological feature |
CN104846054A (en) * | 2015-05-22 | 2015-08-19 | 电子科技大学 | Method for automatically detecting moulds in leucorrhea based on morphological characteristics |
CN105303552A (en) * | 2015-08-07 | 2016-02-03 | 深圳市瀚海基因生物科技有限公司 | Single-molecule positioning device |
CN105303551A (en) * | 2015-08-07 | 2016-02-03 | 深圳市瀚海基因生物科技有限公司 | Single-molecule positioning method |
CN110895800A (en) * | 2015-08-14 | 2020-03-20 | 深圳市真迈生物科技有限公司 | Single-molecule image deviation rectifying method |
CN107492088B (en) * | 2016-06-11 | 2020-12-04 | 青岛华晶生物技术有限公司 | Automatic identification and statistics method for white blood cells in gynecological microscopic image |
CN106295588B (en) * | 2016-08-17 | 2019-10-01 | 电子科技大学 | The automatic identifying method of leucocyte in a kind of leukorrhea micro-image |
CN106875404A (en) * | 2017-01-18 | 2017-06-20 | 宁波摩视光电科技有限公司 | The intelligent identification Method of epithelial cell in a kind of leukorrhea micro-image |
CN109697450B (en) * | 2017-10-20 | 2023-04-07 | 曦医生技股份有限公司 | Cell sorting method |
CN108564114B (en) * | 2018-03-28 | 2022-05-27 | 电子科技大学 | Human body fecal leucocyte automatic identification method based on machine learning |
CN109087320B (en) * | 2018-08-29 | 2022-05-03 | 苏州钮曼精密机电科技有限公司 | Image screening processing method applied to tilt sensor |
CN110415212A (en) * | 2019-06-18 | 2019-11-05 | 平安科技(深圳)有限公司 | Abnormal cell detection method, device and computer readable storage medium |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1653480A (en) * | 2002-05-14 | 2005-08-10 | 维森盖特有限公司 | Optical projection imaging system and method for automatically detecting cells with molecular marker compartmentalization associated with malignancy and disease |
-
2014
- 2014-07-30 CN CN201410369635.9A patent/CN104156951B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1653480A (en) * | 2002-05-14 | 2005-08-10 | 维森盖特有限公司 | Optical projection imaging system and method for automatically detecting cells with molecular marker compartmentalization associated with malignancy and disease |
Non-Patent Citations (2)
Title |
---|
Microbiology and prognostic factors of hospital-and community-acquired aspiration pneumonia in respiratory intensive care unit;Chaojie Wei等;《American Journal of Infection Control》;20131231;880-884 * |
痰液显微图像细胞边缘提取的数学形态学新方法;杨谊等;《计算机应用研究》;20050610(第6期);258-260 * |
Also Published As
Publication number | Publication date |
---|---|
CN104156951A (en) | 2014-11-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104156951B (en) | A kind of white blood cell detection method for BAL fluid smear | |
CN109166095B (en) | Fundus image cup and disc segmentation method based on generation countermeasure mechanism | |
CN109447945B (en) | Quick counting method for basic wheat seedlings based on machine vision and graphic processing | |
CN104794721B (en) | A kind of quick optic disk localization method based on multiple dimensioned spot detection | |
CN109636824A (en) | A kind of multiple target method of counting based on image recognition technology | |
CN104075965B (en) | A kind of micro-image grain graininess measuring method based on watershed segmentation | |
CN106483129B (en) | A method of the leukorrhea trichomonad based on motion estimate detects automatically | |
CN104198355B (en) | Automatic detection method for red cells in feces | |
CN102956035A (en) | Preprocessing method and preprocessing system used for extracting breast regions in mammographic images | |
CN104198497A (en) | Surface defect detection method based on visual saliency map and support vector machine | |
CN104036521A (en) | Novel retina eye fundus image segmenting method | |
CN102024259A (en) | Bacterial colony automatic detection method | |
CN110648330B (en) | Defect detection method for camera glass | |
CN109523524A (en) | A kind of eye fundus image hard exudate detection method based on integrated study | |
Feng et al. | A separating method of adjacent apples based on machine vision and chain code information | |
CN111476804A (en) | Method, device and equipment for efficiently segmenting carrier roller image and storage medium | |
CN110032946A (en) | A kind of aluminium based on machine vision/aluminium blister package tablet identification and localization method | |
CN111126162A (en) | Method, device and storage medium for identifying inflammatory cells in image | |
CN110544262B (en) | Cervical cell image segmentation method based on machine vision | |
Phoulady et al. | An approach for overlapping cell segmentation in multi-layer cervical cell volumes | |
CN106023223A (en) | Orange fruit size describing method and organic fruit size grading method | |
CN111415348B (en) | Method for extracting bubble characteristics in automobile brake pipeline | |
CN106446832B (en) | Video-based pedestrian real-time detection method | |
WO2024016632A1 (en) | Bright spot location method, bright spot location apparatus, electronic device and storage medium | |
CN115272333B (en) | Cup-disk ratio data storage system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant |