CN101807297A - Medical ultrasonic image line detection method - Google Patents
Medical ultrasonic image line detection method Download PDFInfo
- Publication number
- CN101807297A CN101807297A CN200910005674A CN200910005674A CN101807297A CN 101807297 A CN101807297 A CN 101807297A CN 200910005674 A CN200910005674 A CN 200910005674A CN 200910005674 A CN200910005674 A CN 200910005674A CN 101807297 A CN101807297 A CN 101807297A
- Authority
- CN
- China
- Prior art keywords
- straight
- ultrasonic image
- marginal point
- detection method
- medical ultrasonic
- 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
Images
Landscapes
- Ultra Sonic Daignosis Equipment (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a medical ultrasonic image line detection method, which is invented for overcoming the defects of necessary manual operation, bad noiseproof performance, tremendous calculated amount and no real-time detection in the prior art. The method comprises the following steps: (1) the area of interest is chosen; (2) the current frame images are sampled; (3) the marginal point of each sample is found out; and (4) the found marginal point is calculated through random sampling consensus algorithm and a line is chosen to be the detected line according to the calculation result. Through the method, the calculated amount of the medical ultrasonic image line detection processing is reduced greatly, the noiseproof performance is strengthened, noise sensitivity is reduced, and the influence of speckle noise on the detection result is eliminated well and the real-time and automatic detection on the video frequency ultrasonic image can be realized.
Description
Technical field
The present invention relates to a kind of medical ultrasonic image line detection method.
Background technology
The ultrasonic demarcation that straight-line detection in the medical ultrasonic image is widely used in the freedom-arm, three-D ultrasonic image-forming system waits other application.Existing method mainly contains complete manually detect and based on two kinds of methods of straight-line detection of Hough transformation.Complete manual detection method needs operating personnel to select two key points to determine straight line in every two field picture in testing process, and this method is consuming time longer, and whole process all needs manually-operated, and testing result influenced by operating personnel's subjective factor bigger.Can realize the robotization of testing process based on the detection method of Hough transformation, but this method calculated amount is big, can not realize real-time detection to the video ultrasonoscopy, and noise robustness is poor, to the detection poor effect of low SNR images, and testing result can be subjected to having a strong impact on of speckle noise in the ultrasonoscopy.
Summary of the invention
For overcoming above-mentioned defective, the objective of the invention is to propose a kind of method that can detect automatically to the straight line in the medical ultrasonic image of low signal-to-noise ratio, and make the speckle noise in the image as far as possible little to the influence of testing result, and the calculated amount in the processing procedure can not be too big, and computing machine can be handled in real time to the video image that collects.
For achieving the above object, medical ultrasonic image line detection method of the present invention may further comprise the steps:
(1) current two field picture is sampled;
(2) find out the marginal point of each sampling in sample;
(3) use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.
Preferential, also comprise steps A before in above-mentioned step (1): choose the area-of-interest in the image; Corresponding step (1) is specially: the area-of-interest in the current two field picture is sampled.
Preferential, above-mentioned steps A is specially: by clicking the formation convex polygon, the zone in this convex polygon is described area-of-interest in image.
Preferential, above-mentioned steps (1) is specially: choose a plurality of single column of pixels that are in the described area-of-interest in the horizontal direction, each selected single column of pixels is a sampling sample.
Preferential, above-mentioned steps (1) is specially: preestablish a positive integer N as sampling interval, vertically choose the single column of pixels that is in the described area-of-interest as the sample of sampling every N pixel in the horizontal direction then.
Preferential, above-mentioned steps (2) is specially: each sampling sample is carried out difference, find out the marginal point in described each sampling sample.
Preferential, above-mentioned steps (2) is specially: set a gray threshold T1, choose first gray-scale value from top to bottom and surpass the pixel of described gray threshold T1 as marginal point from differentiated described each sampling sample; If all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point.
Preferential, between above-mentioned steps (1) and step (2), also comprise step B: use low-pass filter to carry out filtering to all sampling samples that collect.
Preferential, above-mentioned step (3) is specially:
(31) described each marginal point of finding out is combined into alternative straight in twos;
(32) calculating the consistent of every described alternative straight counts;
(33) the consistent maximal value d that counts of the described alternative straight of search;
(34) if having only consistent the counting of an alternative straight to be d, then this alternative straight is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this alternative straight and, and selected distance and that minimum alternative straight are detected straight line.
Preferential, above-mentioned step (32) is specially:
(321) calculate the distance of each marginal point to every alternative straight;
(322) result and the predefined distance threshold T2 with step (321) compares;
(323) at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.
In the method for the invention described above, reduce calculated amount, and take measure such as filtering, suppress the influence of noise testing result by only the area-of-interest in the image being carried out straight-line detection; And irrelevant information in the rejection image as far as possible, reduce the calculated amount of operation such as Flame Image Process widely, accelerated real-time detection speed to video image.
Description of drawings
Fig. 1 is the FB(flow block) of medical ultrasonic image line detection method embodiment one of the present invention.
Fig. 2 is the FB(flow block) of medical ultrasonic image line detection method embodiment two of the present invention.
Fig. 3 is the FB(flow block) of medical ultrasonic image line detection method embodiment three of the present invention.
Fig. 4 is the synoptic diagram of " setting area-of-interest " step in Fig. 2 or the FB(flow block) shown in Figure 3.
Fig. 5 is the synoptic diagram of Fig. 1 " current two field picture is sampled " step to the FB(flow block) shown in Figure 3.
Fig. 6 is the synoptic diagram of Fig. 1 step of " finding out the marginal point in the sampling sample " to the FB(flow block) shown in Figure 3.
Fig. 7 " uses the random sampling consistency algorithm that the marginal point of finding out is calculated, and chooses straight line as detected straight line according to result of calculation " synoptic diagram of step to FB(flow block) shown in Figure 3 for Fig. 1.
Fig. 8 is the FB(flow block) of medical ultrasonic image line detection method embodiment five of the present invention.
Embodiment
The present invention is further illustrated below in conjunction with drawings and Examples.
Embodiment 1
Medical ultrasonic image line detection method of the present invention as shown in Figure 1, may further comprise the steps:
101, current two field picture is sampled;
Wherein, the process of sampling is such: current frame image was vertically sampled according to the pre-set sampling interval.Preestablish what a positive integer N as sampling interval, vertically choose single column of pixels as a sampling sample every N pixel in the horizontal direction then.Straight-line detection will realize by these samples of sampling.
102, find out the marginal point of sampling in the sample; That is: set a gray threshold T1, will choose pixel that first gray-scale value surpasses described gray threshold T1 after the sampling sample difference from top to bottom, if all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point as marginal point.
103, use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.Its method is: if n bar sampling sample is arranged, n marginal point so just arranged, they can be combined into n* (n-1)/2 alternative straight in twos.To this n* (n-1)/2 alternative straight, we calculate the consistent of them and count: if certain marginal point arrives the distance of certain bar straight line less than predefined distance threshold T2, just think that this marginal point is consistent with this alternative straight; That is to say us at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.By to non-consistent eliminating of putting, can eliminate the influence of speckle noise well to final detection result.The corresponding consistent maximal value d that counts of search n* (n-1)/2 alternative straight.If have only consistent the counting of an alternative straight to be d, then this straight line is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this straight line and, and selected distance and that minimum alternative straight are detected straight line.
In the present embodiment,, can reduce the Flame Image Process amount widely by the current images frame is sampled.Can realize real-time detection to every two field picture in the video.
Embodiment 2
Medical ultrasonic image line detection method of the present invention as shown in Figure 2, may further comprise the steps:
201, set area-of-interest;
That is: before beginning detects in real time, (circle or ellipse belong to a kind of special case of convex polygon by choose polygonal summit formation convex polygon in the image that shows on screen, therefore the convex polygon among the present invention comprises circle or ellipse), zone in the convex polygon is an area-of-interest, and follow-up straight-line detection step will only be carried out the image in the area-of-interest.(as Fig. 4)
202, the area-of-interest in the current two field picture is sampled;
The process of sampling is such: current frame image was vertically sampled according to the pre-set sampling interval.Preestablish what a positive integer N as sampling interval, vertically choose the single column of pixels that is in the area-of-interest as the sampling sample every N pixel in the horizontal direction then.Straight-line detection will realize by these samples of sampling.(as Fig. 5)
203, find out the marginal point of sampling in the sample; That is: set a gray threshold T1, will choose pixel that first gray-scale value surpasses described gray threshold T1 after the sampling sample difference from top to bottom, if all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point as marginal point.(as Fig. 6)
204, use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.Its method is: if n bar sampling sample is arranged, n marginal point so just arranged, they can be combined into n* (n-1)/2 alternative straight in twos.To this n* (n-1)/2 alternative straight, we calculate the consistent of them and count: if certain marginal point arrives the distance of certain bar straight line less than predefined distance threshold T2, just think that this marginal point is consistent with this alternative straight; That is to say us at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.By to non-consistent eliminating of putting, can eliminate the influence of speckle noise well to final detection result.The corresponding consistent maximal value d that counts of search n* (n-1)/2 alternative straight.If have only consistent the counting of an alternative straight to be d, then this straight line is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this straight line and, and selected distance and that minimum alternative straight are detected straight line.(as Fig. 7)
In the present embodiment, by only area-of-interest is carried out straight-line detection reduces calculated amount, and irrelevant information is accelerated the real-time detection speed to every two field picture in the video greatly to straight-line detection result's influence in the rejection image of trying one's best.
Embodiment 3
Medical ultrasonic image line detection method of the present invention as shown in Figure 3, may further comprise the steps:
301, set area-of-interest;
That is: before beginning monitoring in real time, determine a convex polygon by the summit of choosing convex polygon in the image of screen display, the zone in the convex polygon is an area-of-interest, and follow-up straight-line detection step will only be carried out the image in the area-of-interest.(as Fig. 4).
302, the area-of-interest in the current two field picture is sampled;
The process of sampling is such: current frame image was vertically sampled according to the pre-set sampling interval.Preestablish what a positive integer N as sampling interval, vertically choose the single column of pixels that is in the area-of-interest as the sampling sample every N pixel in the horizontal direction then.Straight-line detection will realize by these samples of sampling.(as Fig. 5)
303, use low-pass filter to carry out filtering to all sampling samples that collect;
This step can be suppressed the influence of conventional noise to testing result.
304, find out the marginal point of sampling in the sample.Filtered sampling sample is carried out difference; Set a gray threshold T1, differentiated sampling sample is chosen pixel that first gray-scale value surpasses described gray threshold T1 from top to bottom as marginal point, if all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point.(as Fig. 6)
305, use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.Its method is: if n bar sampling sample is arranged, n marginal point so just arranged, they can be combined into n* (n-1)/2 alternative straight in twos.To this n* (n-1)/2 alternative straight, we calculate the consistent of them and count: if certain marginal point arrives the distance of certain bar straight line less than predefined distance threshold T2, just think that this marginal point is consistent with this alternative straight; That is to say us at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.By to non-consistent eliminating of putting, can eliminate the influence of speckle noise well to final detection result.The corresponding consistent maximal value d that counts of search n* (n-1)/2 alternative straight.If have only consistent the counting of an alternative straight to be d, then this straight line is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this straight line and, and selected distance and that minimum alternative straight are detected straight line.(as Fig. 7)
In the present embodiment, by using low-pass filter to carry out filtering to all sampling samples that collect, thereby reach the purpose of the conventional noise of inhibition to the influence of testing result, strengthened subsequent detection result's accuracy widely, reduce the subsequent image processing amount, can further accelerate real-time detection speed every two field picture in the video.
Embodiment 4
401, set area-of-interest;
That is: before beginning monitoring in real time, determine a convex polygon by the summit of choosing convex polygon in the image of screen display, the zone in the convex polygon is an area-of-interest, and follow-up straight-line detection step will only be carried out the image in the area-of-interest.
402, the area-of-interest in the current two field picture is sampled;
The process of sampling is such: for current frame image, picked at random is in a plurality of single column of pixels in the area-of-interest in the horizontal direction, and each single column of pixels is as a sampling sample.
403, use low-pass filter to carry out filtering to all sampling samples that collect;
This step can be suppressed the influence of the noise of routine to testing result.
404, find out the marginal point of sampling in the sample.Filtered sampling sample is carried out difference; Set a gray threshold T1, differentiated sampling sample is chosen pixel that first gray-scale value surpasses described gray threshold T1 from top to bottom as marginal point, if all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point.(as Fig. 6)
405, use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.Its method is: if n bar sampling sample is arranged, n marginal point so just arranged, they can be combined into n* (n-1)/2 alternative straight in twos.To this n* (n-1)/2 alternative straight, we calculate the consistent of them and count: if certain marginal point arrives the distance of certain bar straight line less than predefined distance threshold T2, just think that this marginal point is consistent with this alternative straight; That is to say us at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.By to non-consistent eliminating of putting, can eliminate the influence of speckle noise well to final detection result.The corresponding consistent maximal value d that counts of search n* (n-1)/2 alternative straight.If have only consistent the counting of an alternative straight to be d, then this straight line is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this straight line and, and selected distance and that minimum alternative straight are detected straight line.(as Fig. 7)
As the distortion of embodiment 3, the step 402 in the present embodiment adopts on-fixed to come at interval area-of-interest is sampled, and also can reach embodiment 3 similar effects.
Embodiment 5
As can be seen, each above-mentioned embodiment can realize the detection to single-frame images from each above-mentioned embodiment.In order to realize that video image is detected in real time, only needing increases the step that the next frame image is detected and gets final product, as shown in Figure 8 after detection of straight lines is finished.
In sum, method of the present invention has overcome prior art needs manually-operated, the very big and shortcoming that can not detect in real time of calculated amount; Strengthened noise immunity, reduced, can eliminate the influence of speckle noise well, can carry out automatic, real-time detection the video ultrasonoscopy to testing result to noise sensitivity.
Obviously, those skilled in the art can carry out various changes and not break away from spiritual scope of the present invention method of the present invention.If therefore these changes belong in claims of the present invention and the equivalent technologies scope thereof, then the present invention also is intended to contain these changes.
Claims (10)
1. a medical ultrasonic image line detection method is characterized in that, may further comprise the steps:
(1) current two field picture is sampled;
(2) find out the marginal point of each sampling in sample;
(3) use the random sampling consistency algorithm that the marginal point of finding out is calculated, and choose straight line as detected straight line according to result of calculation.
2. medical ultrasonic image line detection method as claimed in claim 1 is characterized in that: also comprise steps A before in described step (1): choose the area-of-interest in the image; Corresponding step (1) is specially: the area-of-interest in the current two field picture is sampled.
3. medical ultrasonic image line detection method as claimed in claim 2 is characterized in that: described steps A is specially: by clicking the formation convex polygon, the zone in this convex polygon is described area-of-interest in image.
4. medical ultrasonic image line detection method as claimed in claim 2, it is characterized in that: described step (1) is specially: choose a plurality of single column of pixels that are in the described area-of-interest in the horizontal direction, each selected single column of pixels is a sampling sample.
5. medical ultrasonic image line detection method as claimed in claim 2, it is characterized in that: described step (1) is specially: preestablish a positive integer N as sampling interval, vertically choose the single column of pixels that is in the described area-of-interest as a sampling sample every N pixel in the horizontal direction then.
6. medical ultrasonic image line detection method as claimed in claim 2 is characterized in that: described step (2) is specially: each sampling sample is carried out difference, find out the marginal point in described each sampling sample.
7. medical ultrasonic image line detection method as claimed in claim 6, it is characterized in that: described step (2) is specially: set a gray threshold T1, choose first gray-scale value from top to bottom and surpass the pixel of described gray threshold T1 as marginal point from differentiated described each sampling sample; If all surpass described gray threshold T1 then choose the maximum gradation value point as marginal point.
As claim 1 to the described medical ultrasonic image line detection method of the arbitrary claim of claim 7, it is characterized in that: between step (1) and step (2), also comprise step B: use low-pass filters to carry out filtering all sampling samples that collect.
As claim 1 to the described medical ultrasonic image line detection method of the arbitrary claim of claim 7, it is characterized in that: described step (3) is specially:
(31) described each marginal point of finding out is combined into alternative straight in twos;
(32) calculating the consistent of every described alternative straight counts;
(33) the consistent maximal value d that counts of the described alternative straight of search;
(34) if having only consistent the counting of an alternative straight to be d, then this alternative straight is detected straight line; If it all is d that consistent the counting of many alternative straight arranged, then to every alternative straight wherein all calculate its d the consistent distance of putting this alternative straight and, and selected distance and that minimum alternative straight are detected straight line.
10. medical ultrasonic image line detection method as claimed in claim 9 is characterized in that: described step (32) is specially:
(321) calculate the distance of each marginal point to every alternative straight;
(322) result and the predefined distance threshold T2 with step (321) compares;
(323) at every alternative straight, count on the quantity of the distance of this alternative straight less than the marginal point of described distance threshold T2, this quantity is the consistent of this alternative straight and counts.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910005674XA CN101807297B (en) | 2009-02-16 | 2009-02-16 | Medical ultrasonic image line detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910005674XA CN101807297B (en) | 2009-02-16 | 2009-02-16 | Medical ultrasonic image line detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101807297A true CN101807297A (en) | 2010-08-18 |
| CN101807297B CN101807297B (en) | 2012-07-25 |
Family
ID=42609080
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910005674XA Active CN101807297B (en) | 2009-02-16 | 2009-02-16 | Medical ultrasonic image line detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101807297B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184536A (en) * | 2011-04-19 | 2011-09-14 | 清华大学 | Method and system for extracting straight line and/or line segment end points from image |
| CN102930274A (en) * | 2012-10-19 | 2013-02-13 | 上海交通大学 | Acquisition method and device for medical image |
| CN103610473A (en) * | 2013-11-21 | 2014-03-05 | 海信集团有限公司 | Ultrasonic image adaptive positioning measuring method and device |
| CN105590298A (en) * | 2014-11-07 | 2016-05-18 | 三星电子株式会社 | Extracting and correcting image data of an object from an image |
| CN108322699A (en) * | 2018-01-23 | 2018-07-24 | 东南大学 | A method of based on video images detection arc-flash |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102920436A (en) * | 2012-10-25 | 2013-02-13 | 深圳先进技术研究院 | Processing method of muscle image by utilizing Hough transform |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093543B (en) * | 2007-06-13 | 2010-05-26 | 中兴通讯股份有限公司 | Method for correcting image in 2D code of quick response matrix |
| CN100446037C (en) * | 2007-08-31 | 2008-12-24 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN101315698B (en) * | 2008-06-25 | 2010-06-02 | 中国人民解放军国防科学技术大学 | Characteristic matching method based on straight line characteristic image registration |
| CN101350101B (en) * | 2008-09-09 | 2011-12-07 | 北京航空航天大学 | Method for auto-registration of multi-amplitude deepness image |
-
2009
- 2009-02-16 CN CN200910005674XA patent/CN101807297B/en active Active
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184536A (en) * | 2011-04-19 | 2011-09-14 | 清华大学 | Method and system for extracting straight line and/or line segment end points from image |
| CN102184536B (en) * | 2011-04-19 | 2013-07-10 | 清华大学 | Method and system for extracting straight line and/or line segment end points from image |
| CN102930274A (en) * | 2012-10-19 | 2013-02-13 | 上海交通大学 | Acquisition method and device for medical image |
| CN102930274B (en) * | 2012-10-19 | 2016-02-03 | 上海交通大学 | A kind of acquisition methods of medical image and device |
| CN103610473A (en) * | 2013-11-21 | 2014-03-05 | 海信集团有限公司 | Ultrasonic image adaptive positioning measuring method and device |
| CN105590298A (en) * | 2014-11-07 | 2016-05-18 | 三星电子株式会社 | Extracting and correcting image data of an object from an image |
| CN108322699A (en) * | 2018-01-23 | 2018-07-24 | 东南大学 | A method of based on video images detection arc-flash |
| CN108322699B (en) * | 2018-01-23 | 2020-02-18 | 东南大学 | Method for detecting arc flash based on video image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101807297B (en) | 2012-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101807297B (en) | Medical ultrasonic image line detection method | |
| CN111080582B (en) | Method for detecting defects of inner and outer surfaces of workpiece | |
| CN102175700B (en) | Method for detecting welding seam segmentation and defects of digital X-ray images | |
| CN106303158B (en) | A kind of striped method for detecting abnormality in video image | |
| CN109741314A (en) | A kind of visible detection method and system of part | |
| WO2014136561A1 (en) | Unevenness inspection system, unevenness inspection method, and unevenness inspection program | |
| CN104101600A (en) | Method and apparatus for detecting fine cracks on cross section of continuous casting slab | |
| CN111047655A (en) | High-definition camera cloth defect detection method based on convolutional neural network | |
| JP2001264257A (en) | Image defect detecting device and method, and storage medium storing procedure for image defect detecting method | |
| CN109462999B (en) | Visual inspection method based on learning through data balance and visual inspection device using same | |
| CN106231295A (en) | A kind of display screen image quality automatic checkout system, detection method and application thereof | |
| CN104749801B (en) | High Precision Automatic optical detecting method and system | |
| CN110007493A (en) | Broken bright spot detection method in liquid crystal display | |
| CN115619708A (en) | Method for judging fault based on image recognition of oil quality change of main shaft bearing lubricating oil | |
| CN110880184A (en) | Method and device for carrying out automatic camera inspection based on optical flow field | |
| JP5088165B2 (en) | Defect detection method and defect detection apparatus | |
| CN110376211B (en) | Wet-process-gummed synthetic leather hemming on-line detection device and method | |
| CN108492306A (en) | A kind of X-type Angular Point Extracting Method based on image outline | |
| CN104966283A (en) | Imaging layered registering method | |
| JP2002022671A (en) | Apparatus and method for inspecting inner wall surface of cylinder | |
| CN105067629B (en) | A kind of chromatic circle resistance open defect and resistance value integrated testing method | |
| CN106093052A (en) | A kind of broken yarn detection method | |
| CN104657975B (en) | A kind of method of video image travers Disturbance Detection | |
| CN117237747A (en) | Hardware defect classification and identification method based on artificial intelligence | |
| JP5257063B2 (en) | Defect detection method and defect detection apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |