CN102542282B - Mosaic detection method and mosaic detection device for passive images - Google Patents
Mosaic detection method and mosaic detection device for passive images Download PDFInfo
- Publication number
- CN102542282B CN102542282B CN201010608117XA CN201010608117A CN102542282B CN 102542282 B CN102542282 B CN 102542282B CN 201010608117X A CN201010608117X A CN 201010608117XA CN 201010608117 A CN201010608117 A CN 201010608117A CN 102542282 B CN102542282 B CN 102542282B
- Authority
- CN
- China
- Prior art keywords
- edge map
- binary edge
- template
- mosaic
- image
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a mosaic detection method and a mosaic detection device for passive images, which are applied to the technical field of image detection. The method comprises the following steps of: carrying out edge detection on images to be detected to obtain two-value edge images; carrying out expansion on the two-value edge images by adopting rectangular structure elements; carrying out template matching on the two-value edge images after expansion processing by adopting a plurality of square templates, and overlapping matching points obtained by matching the plurality of square templates and the two-value edge images on one two-value edge image; the two-value edge image formed by overlapping is divided into a plurality of preselection blocks by adopting set thresholds, if the preslection blocks contain a plurality of matching points, one matching point is maintained for forming a template matching point map; carrying out angular-point detection on the images to be detected, wherein if angular points exist in the area coverage of one preselection block, the preselection block is set as the mosaic. For the mosaic detection method and the mosaic detection device provided by the utility model, angular-point detection and template matching are carried out simultaneously, so that the accuracy degree in detecting the mosaic distortion area is improved.
Description
Technical field
The present invention relates to technical field of image detection, relate in particular to a kind of passive images mosaic detection method and device.
Background technology
Along with the development of computer technology, at the Digital Television fault picture, automatically in monitoring, also used widely image recognition technology.Hei Chang, quiet frame, mosaic are the faults that Digital Television often there will be.Block-based method for video coding, due to the mistake in decoding or transmitting procedure, apparent in view blocking effect may appear, namely so-called " mosaic ", the appearance of mosaic has reduced the visual quality of image, how need to detect exactly position that whether every two field picture in video have mosaic, a mosaic, with the exact evaluation of auxiliary video quality for this reason as far as possible.The automatic detection of the Digital Television Hei Chang based on image recognition technology, quiet frame is widely used, and has obtained in practice good effect.Wherein, the detection of mosaic is a difficult point, at present, and the domestic digital television monitoring system that does not also there is the detection of mosaic fault picture.
Meanwhile, scholar both domestic and external and researchist have also proposed some algorithms that detect for mosaic.Great majority still detect for the mosaic in video, can utilize like this front and back frame of present frame with reference to detecting whether contain mosaic.Concrete method comprises:
A, Storey propose as far back as nineteen eighty-three the image block of directly thinking huge with the pixel value difference of front and back frame, are labeled as the mosaic distortion, and Kokaram has improved the method subsequently about 1993, has added motion compensation.
The algorithm that B, domestic scholars have proposed a kind of Schema-based coupling and support vector machine (SVM) in recent years detects the frame that contains mosaic in video sequence automatically.But this method is calculated more complicated, be not easy to the application of extensive video quality evaluation.
C, combine the activity of block boundary luminance difference, local background, distinguish true edge and the texture of blocking effect and image with three restrictive conditions, with the position of detecting more accurately block boundary and the degree of blocking effect.This method can be used for reference the accuracy detected to strengthen the mosaic edges of regions in the mosaic detection to a certain extent, but blocking effect and mosaic or distinguishing, blocking effect detects with ratio of compression much relations, and the effect of finally implementing also can't be estimated.
To sum up analyze, the calculation of complex that in prior art there is the method for image detection, the reference information needed is too much, is not easy to large-scale promotion, and there is defect in some bases on accuracy of detection.
Summary of the invention
The invention provides a kind of passive images mosaic detection method and device, for solving the method calculation of complex of prior art image detection, the reference information needed is too much, is not easy to the problem of large-scale promotion.
The embodiment of the present invention provides a kind of passive images mosaic detection method, comprising:
Treat detected image and carry out rim detection, obtain the binary edge map about described image border to be detected;
Adopt the rectangle structure element to be expanded to binary edge map, make the regular edges of described binary edge map;
Binary edge map after using a plurality of square templates to expansion process carries out template matches, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width binary edge map;
Be divided into a plurality of preliminary election pieces with the threshold value of the setting binary edge map formed that will superpose, if described preliminary election piece comprises a plurality of match points, retain a formation template matches point diagram;
Treat detected image and carry out Corner Detection generation Corner Detection figure;
By template matches point diagram and the Corner Detection figure comparison obtained, if in the regional extent of described arbitrary preliminary election piece, angle point arranged determine that this preliminary election piece is mosaic.
Also provide a kind of passive images mosaic pick-up unit according to said method the present invention, comprising:
Pretreatment unit, carry out rim detection for treating detected image, obtains the binary edge map about described image border to be detected, and adopt the rectangle structure element to be expanded to binary edge map, makes the regular edges of described binary edge map;
Matching unit, carry out template matches for the binary edge map after using a plurality of square templates to expansion process, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width binary edge map; Be divided into a plurality of preliminary election pieces with the threshold value of the setting binary edge map formed that will superpose, if described preliminary election piece comprises a plurality of match points, retain a formation template matches point diagram;
The Corner Detection unit, carry out Corner Detection generation Corner Detection figure for treating detected image;
The mosaic determining unit, for the template matches point diagram by obtaining and Corner Detection figure comparison, if in the regional extent of described arbitrary preliminary election piece, angle point arranged determine that this preliminary election piece is mosaic.
Mosaic detection method and device that the embodiment of the present invention provides, the method that application Corner Detection and template matches are carried out simultaneously, both improved the levels of precision that detects the mosaic distortion zone, can flexible Application on a plurality of application platforms such as image or video.
The accompanying drawing explanation
The process flow diagram that Fig. 1 is a kind of passive images mosaic of embodiment of the present invention detection method;
The schematic diagram that Fig. 2 is image to be detected;
Fig. 3 is the edge binary images that the embodiment of the present invention adopts the canny operator to obtain;
Fig. 4 is the edge binary images after morphological dilations in the real template matches point diagram of the present invention embodiment;
Two class templates that use in the template matches process in Fig. 5 embodiment of the present invention;
Template matches point diagram in Fig. 6 embodiment of the present invention;
Corner Detection figure in Fig. 7 embodiment of the present invention;
A kind of passive images mosaic structure of the detecting device figure in Fig. 8 embodiment of the present invention.
Embodiment
The embodiment of the present invention provides a kind of passive images mosaic detection method, and the method comprises: treat detected image and carry out rim detection, obtain the binary edge map about described image border to be detected; Adopt the rectangle structure element to be expanded to binary edge map, make the regular edges of described binary edge map; Binary edge map after using a plurality of square templates to expansion process carries out template matches, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width figure and forms the template matches point diagram; Described template matches point diagram is divided into to a plurality of preliminary election pieces with the threshold value of setting, if described preliminary election piece comprises a plurality of match points, retains one; Treat detected image and carry out Corner Detection generation Corner Detection figure; By the template matches point diagram that obtains and Corner Detection figure comparison, if in the regional extent of described arbitrary preliminary election piece, angle point is arranged what determine this preliminary election piece is mosaic.
As shown in Figure 1, the embodiment of the present invention provides a kind of passive images mosaic detection method, and concrete steps comprise:
Treat in embodiments of the present invention when detected image is carried out rim detection and can adopt the Canny operator, obtain the binary edge map edge about edge, but be not limited to this kind of detection mode.
Can adopt successively two structural elements [000 in embodiments of the present invention; 111; 000] and [010; 010; 010] image is expanded.After carrying out expansion process, can obtain the relatively accentuated edges image s-edge of rule.
As shown in Figure 5, a plurality of square templates of selecting in the embodiment of the present invention comprise: 1 square template and 4 half square templates.
In concrete applied environment, because MPEG-2 is based on the compaction coding method of piece, therefore always piece or bar of the distorted region of impaired image.Mosaic fault picture in digital TV video frequency just refers to this mistake based on piece or bar.So the size of mosaic is very regular, generally is 16*16, therefore the size of template is elected 16*16 as.
So threshold value 16*16 of described setting in embodiments of the present invention.Wherein, determine whether that a plurality of methods in same preliminary election piece are: the transverse and longitudinal coordinate that adopts point judges whether in same divided by 16 method, and the point finally retained can be the place, the lower right corner at the preliminary election piece.
Adopt in embodiments of the present invention the Harris Corner Detection, wherein the threshold value of Harris angle point response can be made as 20, and this threshold value can be adjusted according to actual needs.
Embodiment of the present invention step 103 is used a plurality of square templates to carry out template matches to described binary edge map, and specific implementation can be:
Each square template is changed into and binary edge map template bianry image of the same size;
Respectively template bianry image and binary edge map are carried out to two-dimension fourier transform;
The correlativity of calculation template bianry image and binary edge map, and obtain the correlativity spectrogram, wherein the implementation for correlation detection can be: by described binary edge map Rotate 180 degree, and the convolutional calculation that postrotational binary edge map and template bianry image are carried out to Fast Fourier Transform (FFT).
According to spectrum peak in described correlativity spectrogram, it is the match point that coordinate is determined square template and image to be detected.
As shown in Figure 8, according to the said method embodiment of the present invention, also provide a kind of passive images mosaic pick-up unit, comprise pretreatment unit 801, matching unit 802, Corner Detection unit 803 and mosaic determining unit 804:
Matching unit 802, carry out template matches for the binary edge map after using a plurality of square templates to expansion process, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width binary edge map; Be divided into a plurality of preliminary election pieces with the threshold value of the setting binary edge map formed that will superpose, if described preliminary election piece comprises a plurality of match points, retain a formation template matches point diagram;
Mosaic determining unit 804, for the template matches point diagram by obtaining and Corner Detection figure comparison, if in the regional extent of described arbitrary preliminary election piece, angle point is arranged what determine this preliminary election piece is mosaic.
A plurality of square templates of described matching unit 802 use carry out template matches to described binary edge map and comprise:
Each square template is changed into and binary edge map template bianry image of the same size;
Respectively template bianry image and binary edge map are carried out to two-dimension fourier transform;
The correlativity of calculation template bianry image and binary edge map, and obtain the correlativity spectrogram;
According to spectrum peak in described correlativity spectrogram, it is the match point that coordinate is determined square template and image to be detected.
The correlativity of described matching unit 802 calculation template bianry images and binary edge map comprises:
By described binary edge map Rotate 180 degree, and the convolutional calculation that postrotational binary edge map and template bianry image are carried out to Fast Fourier Transform (FFT).
Mosaic detection method and device that the application embodiment of the present invention provides, the method that application Corner Detection and template matches are carried out simultaneously, both improved the levels of precision that detects the mosaic distortion zone, can flexible Application on a plurality of application platforms such as image or video.
The present invention passes through features such as the mosaic shape size are regular, grey scale change is obvious, especially use morphological method to strengthen local edge, point after piecemeal processing template coupling is avoided repetition, and edge and angle point, in conjunction with checking mosaic characteristic, have been realized to the detection to mosaic effectively.
Method of the present invention is not limited to the embodiment described in embodiment, and those skilled in the art's technical scheme according to the present invention draws other embodiment, belongs to equally technological innovation scope of the present invention.
Obviously, those skilled in the art can carry out various changes and modification and not break away from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention also is intended to comprise these changes and modification interior.
Claims (8)
1. a passive images mosaic detection method, is characterized in that, comprising:
Treat detected image and carry out rim detection, obtain the binary edge map about described image border to be detected;
Adopt the rectangle structure element to be expanded to binary edge map, make the regular edges of described binary edge map;
Binary edge map after using a plurality of square templates to expansion process carries out template matches, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width binary edge map;
Be divided into a plurality of preliminary election pieces with the threshold value of the setting binary edge map formed that will superpose, if described preliminary election piece comprises a plurality of match points, retain a formation template matches point diagram;
Treat detected image and carry out Corner Detection generation Corner Detection figure;
By template matches point diagram and the Corner Detection figure comparison obtained, if in the regional extent of described arbitrary preliminary election piece, angle point arranged determine that this preliminary election piece is mosaic.
2. the method for claim 1, is characterized in that, described a plurality of square templates comprise 1 square template and 4 half square templates.
3. method as claimed in claim 1 or 2, is characterized in that, described rectangle structure element is [000; 111; 000] and [010; 010; 010].
4. the method for claim 1, is characterized in that, uses a plurality of square templates to carry out template matches to described binary edge map and comprise:
Each square template is changed into and binary edge map template bianry image of the same size;
Respectively template bianry image and binary edge map are carried out to two-dimension fourier transform;
The correlativity of calculation template bianry image and binary edge map, and obtain the correlativity spectrogram;
According to spectrum peak in described correlativity spectrogram, it is the match point that coordinate is determined square template and image to be detected.
5. method as claimed in claim 4, is characterized in that, the correlativity of described calculation template bianry image and binary edge map comprises:
By described binary edge map Rotate 180 degree, and the convolutional calculation that postrotational binary edge map and template bianry image are carried out to Fast Fourier Transform (FFT).
6. a passive images mosaic pick-up unit, is characterized in that, comprising:
Pretreatment unit, carry out rim detection for treating detected image, obtains the binary edge map about described image border to be detected, and adopt the rectangle structure element to be expanded to binary edge map, makes the regular edges of described binary edge map;
Matching unit, carry out template matches for the binary edge map after using a plurality of square templates to expansion process, and the match point that a plurality of square templates and binary edge map coupling is obtained is superimposed upon in a width binary edge map; Be divided into a plurality of preliminary election pieces with the threshold value of the setting binary edge map formed that will superpose, if described preliminary election piece comprises a plurality of match points, retain a formation template matches point diagram;
The Corner Detection unit, carry out Corner Detection generation Corner Detection figure for treating detected image;
The mosaic determining unit, for the template matches point diagram by obtaining and Corner Detection figure comparison, if in the regional extent of described arbitrary preliminary election piece, angle point arranged determine that this preliminary election piece is mosaic.
7. device as claimed in claim 6, is characterized in that, described matching unit is used a plurality of square templates to carry out template matches to described binary edge map to comprise:
Each square template is changed into and binary edge map template bianry image of the same size;
Respectively template bianry image and binary edge map are carried out to two-dimension fourier transform;
The correlativity of calculation template bianry image and binary edge map, and obtain the correlativity spectrogram;
According to spectrum peak in described correlativity spectrogram, it is the match point that coordinate is determined square template and image to be detected.
8. device as claimed in claim 7, is characterized in that, the correlativity of described matching unit calculation template bianry image and binary edge map comprises:
By described binary edge map Rotate 180 degree, and the convolutional calculation that postrotational binary edge map and template bianry image are carried out to Fast Fourier Transform (FFT).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010608117XA CN102542282B (en) | 2010-12-16 | 2010-12-16 | Mosaic detection method and mosaic detection device for passive images |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010608117XA CN102542282B (en) | 2010-12-16 | 2010-12-16 | Mosaic detection method and mosaic detection device for passive images |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102542282A CN102542282A (en) | 2012-07-04 |
| CN102542282B true CN102542282B (en) | 2013-12-04 |
Family
ID=46349134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010608117XA Expired - Fee Related CN102542282B (en) | 2010-12-16 | 2010-12-16 | Mosaic detection method and mosaic detection device for passive images |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102542282B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103473772B (en) * | 2013-09-05 | 2017-04-19 | 北京捷成世纪科技股份有限公司 | Method and device for detecting mosaic image |
| US20170006281A1 (en) * | 2014-05-08 | 2017-01-05 | Huawei Device Co., Ltd. | Video Quality Detection Method and Apparatus |
| CN106372584B (en) * | 2016-08-26 | 2019-06-11 | 浙江银江研究院有限公司 | A kind of video image mosaic detection method |
| CN106447660B (en) * | 2016-09-27 | 2019-01-25 | 百度在线网络技术（北京）有限公司 | Picture detection method and device |
| CN107563970A (en) * | 2017-08-09 | 2018-01-09 | 中国科学院西安光学精密机械研究所 | Method for detecting video image mosaic |
| CN107818568A (en) * | 2017-09-29 | 2018-03-20 | 昆明理工大学 | A kind of video mosaic detection method |
| CN108364282B (en) * | 2018-01-15 | 2022-02-11 | 北京华兴宏视技术发展有限公司 | Image mosaic detection method and image mosaic detection system |
| TWI697871B (en) * | 2019-04-01 | 2020-07-01 | 中華電信股份有限公司 | Inspection system for image containing mosaic and method thereof |
| CN113542864B (en) * | 2020-04-24 | 2023-04-25 | 腾讯科技（深圳）有限公司 | Video splash screen area detection method, device and equipment and readable storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN101236603A (en) * | 2008-02-28 | 2008-08-06 | 上海交通大学 | Vehicle intelligent back vision method |
| CN101551900A (en) * | 2008-03-31 | 2009-10-07 | 新奥特（北京）视频技术有限公司 | Detection method for video mosaic images |
-
2010
- 2010-12-16 CN CN201010608117XA patent/CN102542282B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101110122A (en) * | 2007-08-31 | 2008-01-23 | 北京工业大学 | Large cultural heritage picture pattern split-joint method based on characteristic |
| CN101236603A (en) * | 2008-02-28 | 2008-08-06 | 上海交通大学 | Vehicle intelligent back vision method |
| CN101551900A (en) * | 2008-03-31 | 2009-10-07 | 新奥特（北京）视频技术有限公司 | Detection method for video mosaic images |
Non-Patent Citations (2)
| Title |
|---|
| 司文丽等.马赛克故障图像的分析与检测.《有线电视技术》.2008,(第5期),第113-116页. |
| 马赛克故障图像的分析与检测;司文丽等;《有线电视技术》;20081231(第5期);第113-116页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102542282A (en) | 2012-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102542282B (en) | Mosaic detection method and mosaic detection device for passive images | |
| EP3296953B1 (en) | Method and device for processing depth images | |
| CN106650682A (en) | Method and device for face tracking | |
| WO2016192494A1 (en) | Image processing method and device | |
| CN103026385B (en) | Template is used to switch and feature adaptation provides the method for Object tracking, device and computer program | |
| CN108805863B (en) | Method for detecting image change by combining deep convolutional neural network with morphology | |
| CN107360435B (en) | Blockiness detection methods, block noise filtering method and device | |
| CN109685827B (en) | Target detection and tracking method based on DSP | |
| CN102855617A (en) | Method and system for processing adaptive images | |
| CN103810718B (en) | A kind of strenuous exercise's object detection method and device | |
| CN101018290A (en) | An image processing method and device | |
| CN112347817B (en) | Video target detection and tracking method and device | |
| CN110245600B (en) | Unmanned aerial vehicle road detection method for self-adaptive initial quick stroke width | |
| CN106327464A (en) | Edge detection method | |
| CN111257329A (en) | Smartphone camera defect detection method and detection system | |
| CN110807763A (en) | Method and system for detecting ceramic tile surface bulge | |
| CN103914829A (en) | Method for detecting edge of noisy image | |
| CN114612410A (en) | Novel clothing detects device | |
| CN102118547A (en) | Image weighted filtering method | |
| CN116188379A (en) | Edge defect detection method, device, electronic equipment and storage medium | |
| CN105741326A (en) | Target tracking method for video sequence based on clustering fusion | |
| CN108492306A (en) | A kind of X-type Angular Point Extracting Method based on image outline | |
| Chen et al. | Edge preservation ratio for image sharpness assessment | |
| CN103106663B (en) | Realize the method for SIM card defects detection based on image procossing in computer system | |
| CN103324906B (en) | A kind of method and apparatus of legacy detection |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131204 Termination date: 20191216 |