CN106023137B - A kind of timber method of counting based on contour optimization - Google Patents
A kind of timber method of counting based on contour optimization Download PDFInfo
- Publication number
- CN106023137B CN106023137B CN201610288409.7A CN201610288409A CN106023137B CN 106023137 B CN106023137 B CN 106023137B CN 201610288409 A CN201610288409 A CN 201610288409A CN 106023137 B CN106023137 B CN 106023137B
- Authority
- CN
- China
- Prior art keywords
- image
- timber
- outer profile
- internal periphery
- target
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of timber method of counting based on contour optimization, includes the following steps：Timber image is subjected to size change over；Timber target image is extracted from the image after transformation with color model, and image inward flange information is extracted with boundary operator；The timber target image and marginal information of extraction are merged, pre-segmentation image is obtained；To the profile of pre-segmentation image zooming-out candidate target, and the optimization processing that the outer profile to extracting and Internal periphery are corroded and expanded respectively, it repeats the above process, until the outer profile quantity extracted is constant and there is no until Internal periphery；By above-mentioned optimization process, for the timber candidate target of pre-segmentation image, it can reach removal false-alarm targets and eliminate the purpose of the adhesion between target so that the timber target extracted can be directly used for counting.The verification that the validity of institute's extracting method of the present invention is tested.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to a kind of timber method of counting based on contour optimization.
Background technology
The research of the automatic counting method of timber quantity, to efficiently using forest, control forest excessive and random lumbering has important
Meaning.Traditional method is all the quantity of the timber by manually going statistics felling, this not only consumes a large amount of manpower and materials,
And inefficiency.
The prior art in terms of the automatic counting algorithm of the timber based on image procossing is few；It is existing typical based on figure
As the automatic counting algorithm of timber of processing, all it is to be extracted to wood section with color model, then carries out on this basis
Image pre-segmentation and noise-removed filtering finally again carry these with zone marker scheduling algorithm with extracting effective timber candidate target
The timber target got is counted.However often there are false-alarm targets in the timber candidate target obtained after image pre-segmentation
The problem of many adhesions is mostly had between target, these problems directly influence the precision of timber counting, even result in extraction
To timber target image cannot be used directly for timber count.In the prior art, mathematical morphology operators are mostly used to whole picture figure
False-alarm and adhesion problems after above-mentioned image pre-segmentation are solved as carrying out Global treatment；This method is difficult to design suitable for whole
The universal architecture operator of width image, to cope with various real logs targets and false-alarm targets situation, the complexity between target is viscous
Company also is difficult to effectively handle.
Invention content
For the disadvantages described above or Improvement requirement of the prior art, the present invention provides a kind of timber meter based on contour optimization
Counting method, its object is to improve the precision that timber counting is carried out using image.
To achieve the above object, according to one aspect of the present invention, a kind of timber counting based on contour optimization is provided
Method, this method specifically comprise the following steps：
(1) size change over is carried out to original timber image, obtains new images；
(2) it uses color model to extract timber target area image from the new images, and institute is extracted using boundary operator
State the edge image of new images；
(3) fusion treatment is carried out to the timber target area image and the edge image, obtains pre-segmentation image；
(4) contours optimization process is carried out to the pre-segmentation image, removes the false-alarm targets in pre-segmentation image, and eliminate
Adhesion between target, obtain treated segmentation image；
(5) using image component labeling algorithm, segmentation image carries out timber quantity statistics according to treated.
Preferably, the above-mentioned timber method of counting based on contour optimization, in step (3), according toTimber target area image and edge image are merged；
Wherein, (x, y) refers to the ranks coordinate of pending image pixel, and P (x, y) refers to the wood obtained after pre-segmentation is handled
Material target image, E (x, y) refer to binary edge map, and O (x, y) refers to two-value timber target area extraction image；Using 0 table
Show background pixel, 1 indicates timber target.
Preferably, the above-mentioned timber method of counting based on contour optimization, the contours optimization process in step (4) is one
Iterative process comprising following sub-step：
(4-1) is using the pre-segmentation image as pending image；
(4-2) extracts the Internal periphery and outer profile of pending image；
(4-3) judges whether that Internal periphery is not present, and judges current iteration outer profile quantity and last iteration outer profile number
Whether amount is identical；If so, terminating contours optimization process；If it is not, outside then using current iteration outer profile quantity as last iteration
Outlines, and enter step (4-4)；
(4-4) obtains the area of the outer profile i of pending image；
(4-5) judges whether the area of the outer profile i is less than single sections of timber minimum area, if so, using background
Pixel filling outer profile i, and enter step (4-14)；If it is not, then entering step (4-6)；
(4-6) obtains the circularity of the outer profile i；
(4-7) judges whether the circularity of the outer profile i is less than given roundness threshold, and judges the outer profile i's
Whether area is more than single sections of timber maximum area；If so, outer profile i is carried out corrosion treatment, and enter step (4-
14)；If it is not, then entering step (4-8)；
Wherein, roundness threshold is determined according to the smallest circle angle value of single sections of timber；
(4-8) judges whether the outer profile i has Internal periphery, if so, entering step (4-9)；If it is not, then entering step
Suddenly (4-14)；
(4-9) extracts k-th Internal periphery from the outer profile i；
(4-10) obtains the area of the k-th Internal periphery；
(4-11) judges whether the area of the k-th Internal periphery is less than single sections of timber minimum area, if so, with
Object pixel fills the k-th Internal periphery, and enters step (4-13)；If it is not, then entering step (4-12)；
(4-12) carries out expansion process to the k-th Internal periphery；
(4-13) judges whether the Internal periphery of the outer profile i extracts and finishes, if it is not, then enabling K=K+1, and enters step
(4-9)；If so, entering step (14)；
(4-14) judges whether i reaches current iteration outer profile quantity, if so, this Internal periphery, which is arranged, has label, and
Using the image obtained after expansion process and corrosion treatment as pending image, enter step (4-2)；If it is not, then enabling i
=i+1, and enter step (4-4).
Preferably, the above-mentioned timber method of counting based on contour optimization, profile circularityWherein, S is contoured surface
Product, refers to the number of pixel in profile；L is profile perimeter, as unit of pixel；Circularity indicates the similar journey of timber profile and circle
Degree, maximum value 1；Since the timber profile extracted is frequently not stringent circle, profile circularity C is less than 1；Profile circularity
C shows the contour shape closer to circle closer to 1.
Preferably, the above-mentioned timber method of counting based on contour optimization, corrosion treatment are to replace foreign steamer using background pixel
Wide outermost circle object pixel, to achieve the purpose that reduce the outer profile.
Preferably, the above-mentioned timber method of counting based on contour optimization, expansion process be using background pixel replace close to
The object pixel of the outermost circle of Internal periphery, to achieve the purpose that the expansion Internal periphery.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, can obtain down and show
Beneficial effect：
(1) the timber method of counting provided by the invention based on contour optimization carries out size change over acquisition to timber image
With the timber target image convenient for post-processing；Timber target area image and edge image are extracted to it；And to its carry out
Fusion treatment obtains pre-segmentation image；
Contours optimization process is carried out to pre-segmentation image, the outer profile and Internal periphery that extract are corroded respectively respectively
With the optimization of expansion, until the outer profile quantity extracted is constant and there is no Internal periphery until；Reach removal false-alarm targets and
Eliminate the purpose of the adhesion between target；
For compared with prior art, the present invention is eliminated by contours optimization process between false-alarm targets and target
Adhesion；Since false-alarm targets can cause the mistake of count number to increase, adhesion can make multiple targets be counted as a target,
Therefore, after eliminating these disturbing factors, the precision of timber counting can be improved；
(2) the timber method of counting provided by the invention based on contour optimization in contours optimization process, combines candidate
The circularity and area information of timber objective contour individually implement each profile to corrode or expand according to the different characteristics of profile,
And filling processing；And optimization is iterated to the candidate timber objective contour image that image pre-segmentation obtains, it is empty to reach removal
Alert target and the purpose for eliminating the adhesion between target so that the timber target extracted can be directly used for accurately counting.
Description of the drawings
Fig. 1 is the flow chart for the timber method of counting based on contour optimization that embodiment provides；
Fig. 2 is the flow chart of the contour optimization algorithm in embodiment；
Fig. 3 is the timber image after change of scale in embodiment；
Fig. 4 is the timber target area image extracted in embodiment；
Fig. 5 is the edge image extracted in embodiment；
Fig. 6 is the pre-segmentation image that timber target area image and edge image merge in embodiment；
Fig. 7 is image of the pre-segmentation image Jing Guo the 3rd iteration of contour optimization algorithm in embodiment；
Fig. 8 is to pass through the final segmentation image of contour optimization algorithm in embodiment；
Fig. 9 is the result images that timber counts in embodiment.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
It does not constitute a conflict with each other and can be combined with each other.
Timber method of counting provided in an embodiment of the present invention based on contour optimization, flow is as shown in Figure 1, including as follows
Step：
(1) size change over is carried out to original timber image, obtains new images；
(2) it uses color model to extract timber target area image from the new images, and institute is extracted using boundary operator
State the edge image of new images；
(3) fusion treatment is carried out to the timber target area image and the edge image, obtains pre-segmentation image；
(4) contours optimization process is carried out to the pre-segmentation image, removes the false-alarm targets in pre-segmentation image, and eliminate
Adhesion between target；Obtain treated segmentation image；
(5) using image component labeling algorithm, segmentation image carries out timber quantity statistics to treated.
In the present embodiment, the flow of the contours optimization process of step (4) is as shown in Fig. 2, specific as follows：
(4-1) is using the pre-segmentation image as pending image；
(4-2) extracts the Internal periphery and outer profile of pending image；
(4-3) judges whether that Internal periphery is not present, and judges current iteration outer profile quantity and last iteration outer profile number
Whether amount is identical；If so, terminating contours optimization process；If it is not, outside then using current iteration outer profile quantity as last iteration
Outlines, and enter step (4-4)；In embodiment, the initial value of last iteration outer profile quantity is set as -1, outside current iteration
The initial value of outlines is set as 1；
(4-4) obtains the area of the outer profile i of pending image；
(4-5) judges whether the area of the outer profile i is less than single sections of timber minimum area, if so, using background
Pixel filling outer profile i, and enter step (4-14)；If it is not, then entering step (4-6)；
(4-6) obtains the circularity of the outer profile i；
(4-7) judges whether the circularity of the outer profile i is less than given roundness threshold, and judges the outer profile i's
Whether area is more than single sections of timber maximum area；If so, outer profile i is carried out corrosion treatment, and enter step (4-
14)；If it is not, then entering step (4-8)；
Wherein, roundness threshold is determined according to the smallest circle angle value of single sections of timber；
(4-8) judges whether the outer profile i has Internal periphery, if so, entering step (4-9)；If it is not, then entering step
Suddenly (4-14)；
(4-9) extracts k-th Internal periphery from the outer profile i；
(4-10) obtains the area of the k-th Internal periphery；
(4-11) judges whether the area of the k-th Internal periphery is less than single sections of timber minimum area, if so, with
Object pixel fills the k-th Internal periphery, and enters step (4-13)；If it is not, then entering step (4-12)；
(4-12) carries out expansion process to the k-th Internal periphery；
(4-13) judges whether the Internal periphery of the outer profile i extracts and finishes, if it is not, then enabling K=K+1, and enters step
(4-9)；If so, entering step (14)；
(4-14) judges whether i reaches current iteration outer profile quantity, if so, this Internal periphery, which is arranged, has label, and
Using the image obtained after expansion process and corrosion treatment as new pending image, enter step (4-2)；If it is not,
I=i+1 is then enabled, and enters step (4-4).
Below in conjunction with drawings and the specific embodiments, the timber meter provided by the present invention based on contour optimization is further illustrated
Counting method.
It is the timber image in embodiment after change of scale shown in Fig. 3；The timber image is converted by rgb space
To hsv color space and according to the statistical model based on tone H and saturation degree S pre-established, timber target image is extracted,
Obtain timber target image as shown in Figure 4；In the Fig. 4, bright pixel indicates that edge, dark pixel indicate non-edge；From Fig. 4
As it can be seen that the candidate target of the image there are many false-alarms, has many adhesions between target, cannot be used directly for wood so as to cause it
Material counts.
Using Canny operators from timber image zooming-out shown in Fig. 3 edge, edge image shown in fig. 5 is obtained；It is wherein bright
Pixel indicate edge, dark pixel indicate non-edge.
Timber target image shown in Fig. 4 and edge image shown in fig. 5 are merged, obtaining a web has edge letter
The pre-segmentation image of breath, as shown in Figure 6；Wherein, bright pixel indicates timber target, dark pixels representing background.The image is
Simple pre-segmentation has been done to timber image；As seen from Figure 6, the candidate target of the image still remains that false-alarm is more, between target
Phenomenon more than adhesion causes it to cannot be used directly for timber counting.
Contours optimization process is carried out to above-mentioned pre-segmentation image, remove the false-alarm targets in pre-segmentation image and eliminates target
Between adhesion；In embodiment, the pre-segmentation image obtained after 3 iterative processings is as shown in Figure 7；Wherein, bright pixel
Indicate timber target, dark pixels representing background；From fig.7, it can be seen that the false-alarm of the image has been much less, it is viscous between target
Also it is much less.
Then it is to pass through contours optimization process, the segmentation image finally obtained in embodiment shown in Fig. 8；Wherein, bright pixel
Indicate timber target, dark pixels representing background.As seen from Figure 8, in the image, false-alarm targets have all been eliminated, between target
Adhesion has also removed, which can be directly used for timber counting.
Image is divided to treated using image component labeling algorithm, counts the quantity of profile；Least square fitting is used in combination
The ellipse of one profile completes label and the counting of timber；Fig. 9 is the result images after timber counting in embodiment.
For embodiment, from the timber target area image shown in Fig. 4 extracted it is observed that timber target
There are many adhesions and many false-alarm targets between image, directly go to count using labelling method with this image, error is can be very
Big, because the target of false-alarm will be by error count, and the multiple targets being sticked together will be only counted 1 time；Shown in Fig. 6
Pre-segmentation image can be seen that the adhesion eliminated by marginal portion between target, but adhesion still remains, and false-alarm mesh
Mark is not reduced, and therefore, directly goes to count using labelling method with this image, and error still can be very big；From process shown in Fig. 8
The final segmentation image of contour optimization algorithm can observe that the adhesion and false-alarm targets between target have completely eliminated, at this time
It goes to count using labelling method, it will obtain the correct result of timber number；Fig. 9 illustrates label as a result, every with oval marks
One target, and in the elliptical center sequence that it is counted with digital representation.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, all within the spirits and principles of the present invention made by all any modification, equivalent and improvement etc., should all include
Within protection scope of the present invention.
Claims (5)
1. a kind of timber method of counting based on contour optimization, which is characterized in that include the following steps：
(1) size change over is carried out to original timber image, obtains new images；
(2) it uses color model to extract timber target area image from the new images, and is extracted using boundary operator described new
The edge image of image；
(3) fusion treatment is carried out to the timber target area image and the edge image, obtains pre-segmentation image；
(4) to the pre-segmentation image carry out contours optimization process, remove pre-segmentation image in false-alarm targets, eliminate target it
Between adhesion, obtain treated segmentation image；Including following sub-step：
(4-1) is using the pre-segmentation image as pending image；
(4-2) extracts the Internal periphery and outer profile of pending image；
(4-3) judges whether that Internal periphery is not present, and judges that current iteration outer profile quantity is with last iteration outer profile quantity
It is no identical；If so, terminating contours optimization process；If it is not, then using current iteration outer profile quantity as last iteration outer profile
Quantity, and enter step (4-4)；
(4-4) obtains the area of the outer profile i of pending image；
(4-5) judges whether the area of the outer profile i is less than single sections of timber minimum area, if so, using background pixel
Outer profile i is filled, and enters step (4-14)；If it is not, then entering step (4-6)；
(4-6) obtains the circularity of the outer profile i；
(4-7) judges whether the circularity of the outer profile i is less than given roundness threshold, and judges the area of the outer profile i
Whether single sections of timber maximum area is more than；If so, carrying out corrosion treatment to outer profile i, and enter step (4-14)；If
It is no, then it enters step (4-8)；
Wherein, roundness threshold is determined according to the smallest circle angle value of single sections of timber；
(4-8) judges whether the outer profile i has Internal periphery, if so, entering step (4-9)；If it is not, then entering step
(4-14)；
(4-9) extracts k-th Internal periphery from the outer profile i；
(4-10) obtains the area of the k-th Internal periphery；
(4-11) judges whether the area of the k-th Internal periphery is less than single sections of timber minimum area, if so, using target
K-th Internal periphery described in pixel filling, and enter step (4-13)；If it is not, then entering step (4-12)；
(4-12) carries out expansion process to the k-th Internal periphery；
(4-13) judges whether the Internal periphery of the outer profile i extracts and finishes, if it is not, then enabling K=K+1, and enters step (4-
9)；If so, entering step (4-14)；
(4-14) judges whether i reaches current iteration outer profile quantity, if so, this Internal periphery, which is arranged, has label, and will be through
The image obtained after expansion process and corrosion treatment is crossed as pending image, is entered step (4-2)；If it is not, then enabling i=i+
1, and enter step (4-4)；
(5) image component labeling algorithm is used, segmentation image carries out timber quantity statistics according to treated.
2. timber method of counting as described in claim 1, which is characterized in that in the step (3), according toThe timber target area image and edge image are merged；
Wherein, (x, y) refers to the ranks coordinate of pending image pixel, and P (x, y) refers to the timber mesh obtained after pre-segmentation is handled
Logo image, E (x, y) refer to binary edge map, and O (x, y) refers to two-value timber target area image.
3. timber method of counting as described in claim 1, which is characterized in that profile circularityWherein, S is contoured surface
Product, L are profile perimeter.
4. timber method of counting as described in claim 1, which is characterized in that the corrosion treatment is replaced using background pixel
The outermost circle object pixel of outer profile, to reduce the outer profile.
5. timber method of counting as described in claim 1, which is characterized in that the expansion process is replaced using background pixel
Close to the object pixel of the outermost circle of Internal periphery, to expand the Internal periphery.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610288409.7A CN106023137B (en) | 2016-05-03 | 2016-05-03 | A kind of timber method of counting based on contour optimization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610288409.7A CN106023137B (en) | 2016-05-03 | 2016-05-03 | A kind of timber method of counting based on contour optimization |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106023137A CN106023137A (en) | 2016-10-12 |
| CN106023137B true CN106023137B (en) | 2018-10-16 |
Family
ID=57081728
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610288409.7A Active CN106023137B (en) | 2016-05-03 | 2016-05-03 | A kind of timber method of counting based on contour optimization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106023137B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US12030150B2 (en) | 2019-06-10 | 2024-07-09 | Halliburton Energy Services, Inc. | Drill bit cutter evaluation methods and systems |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106548481A (en) * | 2016-10-18 | 2017-03-29 | 电子科技大学 | A kind of Real-time Collection and processing method of log cross-section information |
| CN110264480B (en) * | 2018-03-12 | 2023-07-07 | 株式会社理光 | Image segmentation method, device and computer readable storage medium |
| CN109977960B (en) * | 2019-04-03 | 2020-02-28 | 杭州深数科技有限公司 | Wood pile information acquisition method, system and device based on neural network |
| WO2020251534A1 (en) * | 2019-06-10 | 2020-12-17 | Halliburton Energy Services, Inc. | Cutter analysis and mapping |
| CN112819745B (en) * | 2019-10-31 | 2023-02-28 | 合肥美亚光电技术股份有限公司 | Nut kernel center worm-eating defect detection method and device |
| CN111462143B (en) * | 2020-03-22 | 2022-12-02 | 华中科技大学 | Watershed algorithm-based insect body recognition and counting method and system |
| CN112116600A (en) * | 2020-08-17 | 2020-12-22 | 内蒙古能建数字信息科技有限公司 | Photovoltaic panel counting method based on image processing |
| CN113160182B (en) * | 2021-04-25 | 2022-04-08 | 湖南九九智能环保股份有限公司 | Stacking section counting system based on image recognition |
| CN113496490B (en) * | 2021-09-06 | 2021-11-09 | 南通弈驰新型建材科技有限公司 | Wood board surface defect detection method and system based on computer vision |
| CN113714662B (en) * | 2021-09-06 | 2022-06-24 | 玖陆零信安创(盐城)科技有限公司 | Cutting size adjusting mechanism for intelligent household wood board cutting equipment |
| CN115937549B (en) * | 2023-01-09 | 2023-08-04 | 肇庆学院 | Wood quantity counting method based on color and shape characteristics |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091180A (en) * | 2014-07-14 | 2014-10-08 | 金陵科技学院 | Method for recognizing trees and buildings in outdoor scene image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004129750A (en) * | 2002-10-09 | 2004-04-30 | Ishimoku:Kk | Decorative sheet |
-
2016
- 2016-05-03 CN CN201610288409.7A patent/CN106023137B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104091180A (en) * | 2014-07-14 | 2014-10-08 | 金陵科技学院 | Method for recognizing trees and buildings in outdoor scene image |
Non-Patent Citations (2)
| Title |
|---|
| Holocene variations of wildfire occurrence as a guide for sustainable management of the northeastern Canadian boreal forest;Ahmed El-Guellab等;《Forest Ecosystems》;20150915(第03期);第1-7页 * |
| 图像处理在木材材积计算中的应用研究;李明涛;《中国优秀硕士学位论文全文数据库信息科技辑》;20080515(第05期);摘要，第5-34页 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US12030150B2 (en) | 2019-06-10 | 2024-07-09 | Halliburton Energy Services, Inc. | Drill bit cutter evaluation methods and systems |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106023137A (en) | 2016-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106023137B (en) | A kind of timber method of counting based on contour optimization | |
| CN111243032B (en) | Full-automatic detection method for checkerboard corner points | |
| CN108776140B (en) | Machine vision-based printed matter flaw detection method and system | |
| JP4172941B2 (en) | Land parcel data creation method and apparatus | |
| CN109447945B (en) | Quick counting method for basic wheat seedlings based on machine vision and graphic processing | |
| CN108364280A (en) | Structural cracks automation describes and width accurately measures method and apparatus | |
| CN104851085B (en) | The automatic method and system for obtaining detection zone in image | |
| CN103942797B (en) | Scene image text detection method and system based on histogram and super-pixels | |
| CN107203990A (en) | A kind of labeling damage testing method based on template matches and image quality measure | |
| CN106709500B (en) | Image feature matching method | |
| CN110119741A (en) | A kind of card card image information recognition methods having powerful connections | |
| CN104680161A (en) | Digit recognition method for identification cards | |
| CN106709484B (en) | Digital identification method of digital instrument | |
| CN109815822B (en) | Patrol diagram part target identification method based on generalized Hough transformation | |
| CN104112132A (en) | Automatic gun number identification method | |
| CN107389693A (en) | A kind of printed matter defect automatic testing method based on machine vision | |
| CN107609558A (en) | Character image processing method and processing device | |
| CN115439476B (en) | Silk-screen defect detection method and device based on image analysis | |
| CN110674812B (en) | Civil license plate positioning and character segmentation method facing complex background | |
| CN106815583A (en) | A kind of vehicle at night license plate locating method being combined based on MSER and SWT | |
| CN104077775A (en) | Shape matching method and device combined with framework feature points and shape contexts | |
| CN111161295A (en) | Background stripping method for dish image | |
| CN110176020A (en) | A kind of bird's nest impurity method for sorting merging 2D and 3D rendering | |
| CN109635679B (en) | Real-time target paper positioning and loop line identification method | |
| CN106599891A (en) | Remote sensing image region-of-interest rapid extraction method based on scale phase spectrum saliency |
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 | ||
| GR01 | Patent grant |