CN107590805A - Trichomonad detection method based on motion vector - Google Patents
Trichomonad detection method based on motion vector Download PDFInfo
- Publication number
- CN107590805A CN107590805A CN201710839719.8A CN201710839719A CN107590805A CN 107590805 A CN107590805 A CN 107590805A CN 201710839719 A CN201710839719 A CN 201710839719A CN 107590805 A CN107590805 A CN 107590805A
- Authority
- CN
- China
- Prior art keywords
- trichomonad
- image
- frame difference
- motion vector
- detection
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of trichomonad detection method based on motion vector.The detection method includes：Trichomonad sample is gathered, with the detection samples pictures under professional CCD camera continuous acquisition high power lens, is automatically analyzed to obtain by computer and schemed for moving the series of tests of trichomonad detection, the sample image as training；Then it is combined to obtaining serial picture progress background model with frame difference method to obtain trichomonad target；Finally detected according to the image characteristics extraction of trichomonad.The present invention realizes the automation and intellectuality of detection by the trichomonad detection method based on motion vector, it is ensured that the stability of Detection results improves detection efficiency and accuracy rate when great amount of samples detects, and has very strong practicality and application value.
Description
Technical field
The present invention relates to image processing field, refers to a kind of trichomonad detection method based on motion vector.
Technical background
Trichomonad detection is exactly that detection is identified using image recognition principle and correlation technique according to the characteristic information of trichomonad
Process.
Trichomonad is a kind of atomic small protozoon biology amphitrichous, can not with the naked eye see, must observe under the microscope.Vagina
Trichomonad long-term existence intravaginal can damage cervix, cause cervix inflammation symptom repeatedly.Trichomonas vaginitis is a kind of compare
Common gynaecological imflammation, the serious health that threaten women of trichomonas vaginitis.The method of current detection trichomonad is mainly had illicit sexual relations
Color method, cultivation, immunological method etc..The above method is carried out by manually being detected by medical science, not over intellectuality
Processing, realizes the detection of efficiently and accurately, in existing detection technique, under operating efficiency, accuracy rate is not high, enterprising to medical science
Large effect be present in the diagnosis of one step.And the detection for moving target mainly has frame difference method at present, background subtraction method, optical flow method
Deng.Frame difference hair is used alone for trichomonad sample image sequence, more interference around trichomonad target area be present, this method for
Chaff interference is sensitive, and robustness is relatively low, and target detection is inaccurate.
Therefore, for drawbacks described above, prior art could be improved and develop.
The content of the invention
The technical problem to be solved in the present invention is, for the drawbacks described above of prior art, there is provided one kind is based on motion arrow
The trichomonad detection method of amount, it is intended to passed through by intelligent processing method by detecting sample under professional CCD camera continuous acquisition high power lens
Picture, application background model are combined extraction trichomonad target with frame difference method, automatically analyze sequence of pictures by computer and detect trichomonad.
The automation and intellectuality of detection are realized, Detection accuracy, efficiency and the stability of detection is effectively ensured.
The technical proposal for solving the technical problem of the invention is as follows：
A kind of trichomonad detection method based on motion vector, wherein, including：Step A, by professional CCD camera continuous acquisition
Samples pictures are detected under high power lens, is automatically analyzed to obtain by computer and schemed for moving the series of tests of trichomonad detection, as instruction
Experienced sample image；Step B, background model is carried out with the processing of frame difference to obtain trichomonad target to obtaining serial picture；Step C,
Detected according to the image characteristics extraction of trichomonad.
The trichomonad detection method based on motion vector, wherein, trichomonad is a kind of protozoon biology, and its profile is justified for class
Shape, appearance profile are more smooth；It is live body in the sample, meeting autokinetic movement, has the change of position.Want acquisition trichomonad to exist
It must be the continuous image sequence of trichomonad in the same context that characteristics of image in image adopts figure using CCD camera.
The trichomonad detection method based on motion vector, wherein, the maximum feature of trichomonad is its motion feature, for drop
Frame difference method is used alone in worm sample image sequence, more interference around trichomonad target area be present, and this method is quick for chaff interference
Sense, robustness are relatively low.And sample image is analyzed, except trichomonad and small impurity have movement, other material sites are fixed, liquid
The gray scale of body is stable, can establish effective background model.
The trichomonad detection method based on motion vector, wherein, the acquisition of target trichomonad is to use Background Modeling
With the method that frame difference method is combined, specific steps include：
B1 background model) is established, specific fusion treatment is carried out using a few frame original images of continuous acquisition as needed, obtains
To background image；
B2 background) is reduced by artwork, obtains foreground image F；
B3) adjacent artwork progress frame is poor, obtains frame difference image diff2；
B4) foreground picture and frame difference figure are combined, obtain new target image M.The trichomonad detection based on motion vector
Method, wherein, establishing background model specific steps includes：
B5) initial background image Backing；Using the first two field picture as the initialisation image of background, formula is initialized
It is as follows:
Backimg (x, y)=O0(x, y)
B6) adjacent artwork carries out frame difference processing, obtains frame difference figure；
The frame difference figure of adjacent two field picture is calculated, consecutive frame same point variation of image grayscale scope, which exceedes, specifies size T, then
The gray scale is 255 in frame difference figure, it is on the contrary then for 0, thus obtain binaryzation frame difference and scheme diff, calculation formula is as follows；
B7 background image) is established according to frame difference figure information；
The gray scale of every bit is superimposed by each frame artwork and obtained in background image, for any point on Background, if i-th
The upper point of subframe difference figure is black, then the point of the i-th secondary artwork participates in superposition, to participate in the institute of superposition a little with initial background figure
The point seeks its gray average, that is, obtains the gray scale of the point in Background.Calculation formula is as follows：
M is to meet diffiThe number of (x, y)=0
The trichomonad detection method based on motion vector, wherein, step B2 is specifically included：Prospect according to its gray scale whether
It is divided into two classes higher than background image gray scale：One kind is high brightness prospect, i.e. prospect gray scale is higher than background image, such calculation formula
It is as follows：
One kind is low-light level prospect, i.e. the gray scale of prospect is less than background image, and such calculation formula is as follows：
The trichomonad detection method based on motion vector, wherein, step B3 is specifically included：According in front and rear two frames figure,
It is also two classes that frame difference figure is divided to by the situation of change of gray scale：The first kind is high luminance targets frame difference image, i.e., target area is adjacent two
By secretly brightening in two field picture, such calculation formula is as follows：
It is low luminance target frame difference image in the second class, it is with the first kind on the contrary, i.e. target area is in adjacent two field pictures
By bright dimmed, such calculation formula is as follows：
The trichomonad detection method based on motion vector, wherein, step B4 is specifically included：It is preceding when being combined computing
Scape figure is combined computing with target frame difference figure according to its brightness height.I.e. first kind foreground picture is transported with first kind frame difference figure
Calculate, i.e. the second class foreground picture and the second class frame difference figure carry out computing.Its operation rule is as follows：
Mi(x, y)=Fi(x, y) ＆diff2i(x, y) 1≤；≤n
The trichomonad detection method based on motion vector, wherein, the detection for trichomonad disturbs because liquid is present, and removes
Outside target trichomonad can move, impurity of other small volumes etc. can also produce movement, to go deimpurity influence, it is necessary to target
After figure carries out graphics opening and closing operation, chosen according to the area of each isolated area, remove noise and impurity.
The trichomonad detection method based on motion vector, wherein, for the recognition detection of trichomonad, it is necessary to special according to trichomonad
Some features identify, so first to analyze the feature for obtaining trichomonad, specially trichomonad profile class circle, contour smoothing；In motion
The motion amplitude of trichomonad is bigger than the mobile range of other materials, and trichomonad profile is more smooth than other impurities；The movement of object is except presence
Horizontal movement, also exist and move up and down, and the material moved up and down its gray scale has larger bright dark change；Liquid is in the picture
Gray scale it is stable.Detection is identified to trichomonad according to these features of trichomonad, then counts trichomonad target numbers, and target
The number of image frames and image coordinate at place.
The detection method of the present invention is easy, quick, using intelligent and automatic business processing mode, with image procossing skill
Art, the detection speed and efficiency and accuracy rate of trichomonad are highly desirable improved, it is low to avoid the slow efficiency of traditional detection speed
The drawbacks of, foundation can be provided for further medical diagnosis.
The trichomonad target of motion can be gone out with effective detection using the detection method of the present invention, it is ensured that great amount of samples detects
When Detection results stability and its detection efficiency, there is very strong practicality and application value.
Brief description of the drawings
Fig. 1 is the flow chart of the preferred embodiment of the trichomonad method of inspection of the invention based on motion vector.
Fig. 2 is a kind of picture of trichomonad.
Fig. 3-7 is continuous picture sequence, the trichomonad object in figure in white circular to find.
Fig. 8 is using context of methods, randomly selects sequence of pictures and is detected, statistical result.
Embodiment
To make the objects, technical solutions and advantages of the present invention clearer, clear and definite, develop simultaneously embodiment pair referring to the drawings
The present invention is further described.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and do not have to
To limit the present invention.
Referring to Fig. 1, Fig. 1 is the flow chart of the trichomonad detection method preferred embodiment of the invention based on motion vector.
As shown in figure 1, the embodiment of the present invention provides a kind of trichomonad detection method based on motion vector, including following step
Suddenly：
S100, by detecting samples pictures under professional CCD camera continuous acquisition high power lens, automatically analyzing acquisition by computer is used for
Move the series of tests figure of trichomonad detection, the sample image as training；
Trichomonad as shown in Figure 2 is a kind of protozoon biology, and its profile is similar round, and appearance profile is more smooth, and it is in sample
In be live body, can autokinetic movement, have the change of position.
If Fig. 3-7 is consecutive image sequence, it is desirable to obtain the characteristics of image of trichomonad in the picture and use CCD camera Cai Tubi
It must be the continuous image sequence of trichomonad in the same context.The purpose that the present invention carries out trichomonad detection is exactly the drop in Fig. 2
Detection is identified in worm, and so as to be separated with other cell impurities, foundation is provided for medical diagnosis.
S200, background model is carried out with the processing of frame difference to obtain trichomonad target to obtaining serial picture；
In the present invention, the maximum feature of trichomonad is its motion feature, and it is poor that frame is used alone for trichomonad sample image sequence
Method, more interference around trichomonad target area be present, this method is sensitive for chaff interference, and robustness is relatively low.And to sample image
Analyzed, except trichomonad and small impurity have movement, other material sites are fixed, and the gray scale of liquid is stable, can establish effectively
Background model, so the acquisition of target trichomonad is the method being combined with frame difference method using Background Modeling, specific steps
Including：
1. establishing background model, specific fusion treatment is carried out using 6 frame original images of continuous acquisition as needed, is obtained
Background image.Specially：1) initial background image Backing；Using the first two field picture as the initialisation image of background, initially
Change formula：
Backimg (x, y)=O0(x, y)
2) adjacent artwork carries out frame difference processing, obtains frame difference figure；Calculate the frame difference figure of adjacent two field picture, consecutive frame same point
Locate variation of image grayscale scope and exceed to specify size T, then on the contrary the gray scale is 255 in frame difference figure, then be 0, thus obtains two
Value frame difference figure diff, calculation formula are as follows：
3) background image is established according to frame difference figure information；The gray scale of every bit is superimposed by each frame artwork and obtained in background image
Arrive, for any point on Background, if the upper point of the i-th subframe difference figure is black, the point of the i-th secondary artwork participates in superposition, right
A little with initial background figure, the point seeks its gray average for the institute of participation superposition, that is, obtains the gray scale of the point in Background.Calculate
Formula is as follows：
M is to meet diffi(x, y)=0
Number
2. reducing background by artwork, foreground image F is obtained.Whether prospect is divided into higher than background image gray scale according to its gray scale
Two classes：One kind is high brightness prospect, i.e. prospect gray scale is higher than background image, and formula is as follows：
One kind is low-light level prospect, i.e. the gray scale of prospect is less than background image, and such calculation formula is as follows：
3. adjacent artwork progress frame is poor, frame difference image diff2 is obtained.According in front and rear two frames figure, the situation of change of gray scale
It is also two classes that frame difference figure, which is divided to,：The first kind is high luminance targets frame difference image, i.e., target area in adjacent two field pictures by secretly becoming
Bright, such calculation formula is as follows：
Second class is low luminance target frame difference image, its with the first kind on the contrary, i.e. target area in adjacent two field pictures by
It is bright dimmed, such calculation formula：
4. foreground picture and frame difference figure are combined, new target image M is obtained.When being combined computing, foreground picture and target
Frame difference figure is combined computing according to its brightness height.That is first kind foreground picture and the difference figure progress computing of first kind frame, i.e., second
Class foreground picture and the second class frame difference figure carry out computing.Operation rule is that pixel correspondingly seeks ＆, and the operation rule is as follows：
Mi(x, y)=Fi(x, y) ＆diff2i(x, y) 1≤i≤n
S300, detected according to the image characteristics extraction of trichomonad.
In the present invention, detection for trichomonad disturbs because liquid is present, in addition to target trichomonad can move, other volumes compared with
Small impurity etc. can also produce movement, after going deimpurity influence, it is necessary to carry out graphics opening and closing operation to target figure, according to
The area of each isolated area is chosen, and removes noise and impurity.
In the present invention, for the recognition detection of trichomonad, it is necessary to identified according to the distinctive feature of trichomonad, so first to analyze
Obtain the feature of trichomonad, specially trichomonad profile class circle, contour smoothing；Shifting of the motion amplitude of trichomonad than other materials in motion
Dynamic amplitude is big, and trichomonad profile is more smooth than other impurities；The movement of object also exists and moved up and down except horizontal movement be present, and on
Its gray scale of the material of lower movement has larger bright dark change；The gray scale of liquid in the picture is stable.According to these spies of trichomonad
Sign trichomonad is identified detection, then counts trichomonad target numbers, and number of image frames and image coordinate where target.
If Fig. 8 is the obtained statistical result using this method.The present invention passes through the trichomonad based on motion vector in summary
Detection method, realize the automation and intellectuality of detection, it is ensured that the stability of Detection results carries when great amount of samples detects
High detection efficiency and accuracy rate, there is very strong practicality and application value.
It should be appreciated that the application of the present invention is not limited to above-mentioned citing, for those skilled in the art,
It can according to the above description be improved or be converted, all these modifications and variations should all belong to appended claims of the present invention
Protection domain.
Claims (9)
- A kind of 1. trichomonad detection method based on motion vector, it is characterised in that including：Step A, trichomonad sample is gathered, with the detection samples pictures under professional CCD camera continuous acquisition high power lens；Step B, background model is carried out with frame difference method processing to obtain trichomonad target to obtaining serial picture；Step C, detected according to the image characteristics extraction of trichomonad, realize the detection to the target image.
- 2. the detection method of the trichomonad according to claim 1 based on motion vector, it is characterised in that enter to sample image There is movement in row analysis, the sample image, other materials position is fixed except trichomonad and small impurity, and the gray scale of liquid is stable, builds Found effective background model.
- 3. the detection method of the trichomonad according to claim 2 based on motion vector, it is characterised in that for trichomonad target Thing has greater brightness to change this feature when moving up and down, and object is divided into bright and dark two class is detected.
- 4. the detection method of the trichomonad according to claim 1 based on motion vector, it is characterised in that enter for target figure After row graphics opening and closing operation, chosen according to the area of each isolated area, remove noise and impurity.
- 5. the detection method of the trichomonad according to claim 1 based on motion vector, it is characterised in that the step B tools Body includes：B1 background model) is established；B2 background) is reduced by artwork, obtains foreground image；B3) adjacent artwork progress frame is poor, obtains frame difference image；B4) foreground picture and frame difference figure are combined, obtain new target image.
- 6. the detection method of the trichomonad according to claim 4 based on motion vector, it is characterised in that the step B1 tools Body includes：B5) initial background image, initialization formula are as follows：Backimg (x, y)=Oo(x, y)B6) adjacent artwork carries out frame difference processing, obtains frame difference figure；The frame difference figure of adjacent two field picture is calculated, consecutive frame same point variation of image grayscale scope, which exceedes, specifies size T, then frame is poor This in figure gray scale is 255, it is on the contrary then for 0, thus obtain binaryzation frame difference and scheme Diff, calculation formula is as follows：B7 background image) is established according to frame difference figure information；The gray scale of every bit is superimposed by each frame artwork and obtained in background image.For any point on Background, if The upper point of i subframes difference figure is black, then the point of the i-th secondary artwork participates in superposition.To participate in superposition institute a little andThe gray scale of point.Calculation formula is as followsM is to meet diffiThe number of (x, y)=0.
- 7. the detection method of the trichomonad according to claim 4 based on motion vector, it is characterised in that the step B2 tools Body includes：Whether prospect higher than background image gray scale is divided into two classes according to its gray scale：One kind is high brightness prospect, i.e. prospect gray scale It is as follows higher than background image, calculation formula：One kind is low-light level prospect, i.e. the gray scale of prospect is less than background image, and such calculation formula is as follows：。
- 8. the detection method of the trichomonad according to claim 4 based on motion vector, it is characterised in that the step B3 tools Body includes：According in front and rear two frames figure, it is also two classes that frame difference figure is divided to by the situation of change of gray scale：The first kind is high luminance targets frame By secretly brightening in adjacent two field pictures, such calculation formula is as follows for difference image, i.e. target area：Second class is low luminance target frame difference image, its with the first kind on the contrary, i.e. target area in adjacent two field pictures by bright change Secretly, such calculation formula is as follows：。
- 9. the detection method of the trichomonad according to claim 4 based on motion vector, it is characterised in that the step B4 tools Body：When being combined computing, foreground picture is combined computing with target frame difference figure according to its brightness height.That is first kind foreground picture Computing is carried out with first kind frame difference figure, i.e. the second class foreground picture and the second class frame difference figure carry out computing.Its operation rule is as follows：Mi(x, y)=Fi(x, y) ＆diff2i(x, y) 1≤i≤n.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710839719.8A CN107590805A (en) | 2017-09-15 | 2017-09-15 | Trichomonad detection method based on motion vector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710839719.8A CN107590805A (en) | 2017-09-15 | 2017-09-15 | Trichomonad detection method based on motion vector |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107590805A true CN107590805A (en) | 2018-01-16 |
Family
ID=61047141
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710839719.8A Pending CN107590805A (en) | 2017-09-15 | 2017-09-15 | Trichomonad detection method based on motion vector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107590805A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931754A (en) * | 2020-10-14 | 2020-11-13 | 深圳市瑞图生物技术有限公司 | Method and system for identifying target object in sample and readable storage medium |
| CN113450354A (en) * | 2021-08-30 | 2021-09-28 | 山东仕达思生物产业有限公司 | Trichomonas detection method based on convolutional neural network and swing activity characteristics |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102713914A (en) * | 2009-10-19 | 2012-10-03 | 提拉诺斯公司 | Integrated health data capture and analysis system |
-
2017
- 2017-09-15 CN CN201710839719.8A patent/CN107590805A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102713914A (en) * | 2009-10-19 | 2012-10-03 | 提拉诺斯公司 | Integrated health data capture and analysis system |
Non-Patent Citations (5)
| Title |
|---|
| 段玉瑶等: ""养殖场环境下的生猪多目标检测与跟踪方法"", 《光学学报》 * |
| 马丽文: ""基于纹理分析的妇科显微图像的成分统计与分析"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 高凯亮等: ""一种基于帧差法与背景减法的运动目标检测新方法"", 《电讯技术》 * |
| 高哲: ""运动目标检测与跟踪算法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 黄超: ""基于智能视觉的机场区域移动目标的检测方法研究"", 《现代电子技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111931754A (en) * | 2020-10-14 | 2020-11-13 | 深圳市瑞图生物技术有限公司 | Method and system for identifying target object in sample and readable storage medium |
| CN111931754B (en) * | 2020-10-14 | 2021-01-15 | 深圳市瑞图生物技术有限公司 | Method and system for identifying target object in sample and readable storage medium |
| CN113450354A (en) * | 2021-08-30 | 2021-09-28 | 山东仕达思生物产业有限公司 | Trichomonas detection method based on convolutional neural network and swing activity characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108827316B (en) | Mobile robot visual positioning method based on improved Apriltag | |
| CN108416766B (en) | Double-side light-entering type light guide plate defect visual detection method | |
| CN104715238B (en) | A kind of pedestrian detection method based on multi-feature fusion | |
| CN107678192B (en) | Mura defect detection method based on machine vision | |
| US20180342078A1 (en) | Information processing device, information processing method, and information processing system | |
| CN104198497B (en) | Surface defect detection method based on visual saliency map and support vector machine | |
| CN104101294B (en) | Analytical equipment, analysis program and analysis system | |
| CN108470338B (en) | A kind of water level monitoring method | |
| CN106886216A (en) | Robot automatic tracking method and system based on RGBD Face datections | |
| WO2017120796A1 (en) | Pavement distress detection method and apparatus, and electronic device | |
| CN108521554A (en) | Large scene multi-target cooperative tracking method, intelligent monitor system, traffic system | |
| US9280209B2 (en) | Method for generating 3D coordinates and mobile terminal for generating 3D coordinates | |
| CN105678737B (en) | A kind of digital picture angular-point detection method based on Radon transformation | |
| CN109182081A (en) | A kind of unicellular separation system based on image processing model | |
| WO2022267300A1 (en) | Method and system for automatically extracting target area in image, and storage medium | |
| CN107590805A (en) | Trichomonad detection method based on motion vector | |
| Shah et al. | Automatic detection and classification of tuberculosis bacilli from camera-enabled Smartphone microscopic images | |
| CN106529441A (en) | Fuzzy boundary fragmentation-based depth motion map human body action recognition method | |
| Nakamichi et al. | Classification of circulating tumor cells in fluorescence microscopy images based on SqueezeNet | |
| CN108765463A (en) | A kind of moving target detecting method calmodulin binding domain CaM extraction and improve textural characteristics | |
| CN105761244A (en) | Method, device and system for determining nematode mortality rate | |
| CN108428218A (en) | A kind of image processing method of removal newton halation | |
| CN103632373B (en) | A kind of flco detection method of three-frame difference high-order statistic combination OTSU algorithms | |
| CN113450354B (en) | Trichomonas detection method based on convolutional neural network and swing activity characteristics | |
| CN105184244B (en) | Video human face detection method and device |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180116 |
|
| RJ01 | Rejection of invention patent application after publication |