CN102222325A - Method for detecting object contour based on phase characteristic - Google Patents
Method for detecting object contour based on phase characteristic Download PDFInfo
- Publication number
- CN102222325A CN102222325A CN 201110163800 CN201110163800A CN102222325A CN 102222325 A CN102222325 A CN 102222325A CN 201110163800 CN201110163800 CN 201110163800 CN 201110163800 A CN201110163800 A CN 201110163800A CN 102222325 A CN102222325 A CN 102222325A
- Authority
- CN
- China
- Prior art keywords
- contour
- characteristic pattern
- image
- receptive field
- classical receptive
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to the technical field of machine vision and discloses a method for detecting an object contour based on a phase characteristic, which is disclosed specific to the problem of difficulty in extracting a contour from a position with low image contrast existing in a method for detecting a non-classical receptive field suppression contour. The method specifically comprises the following steps of: filtering; drawing a characteristic pattern; drawing a characteristic pattern which is subjected to edge preserving smooth filtering; and drawing a contour image which is subjected to non-classical receptive field suppression and performing binaryzation. In the contour detection method, the integrity of the characteristic is kept by using the phase characteristic of an image as well as the stability and importance of image phase information, so that contour detection is not influenced by the inconsistence of image brightness or the change of contrast; and moreover, a target contour is effectively distinguished from a texture edge in combination with a non-classical receptive field characteristic in a human vision mechanism, and textures are suppressed.
Description
Technical field
The invention belongs to the machine vision technique field, particularly a kind of detection method of contour of object.
Background technology
Human vision system can be according to the variation of visual environment and condition, image information to input is carried out the adjustment and the processing of various subjectivities, up to the present, more representational profile testing method based on vision mechanism is that non-classical receptive field suppresses profile testing method, specifically can be referring to document: Grigorescu C, Petkov N, Westenberg M, Contour detection based on nonclassical receptive field inhibition, IEEE Transactions on Image Processing, vol.12, no.7,729～739,2003.This method has utilized the rejection characteristic of the neuronic non-classical receptive field of primary visual cortex to suppress background texture information, and then reaches the purpose of outstanding target object profile, comprising isotropy inhibition method and anisotropy inhibition method.This profile of obtaining based on the method for vision mechanism detects effect generally all significantly better than traditional method, but the deficiency that exists is to be subject to the influence that the input picture contrast changes, and the profile at image low contrast place is difficult to extract.
Summary of the invention
The objective of the invention is to suppress the problem that the profile of profile testing method at image low contrast place is difficult to extract, proposed a kind of contour of object detection method based on phase propetry in order to solve existing non-classical receptive field.
To achieve these goals, technical scheme of the present invention is: a kind of contour of object detection method based on phase propetry comprises the steps:
S1. Filtering Processing is specially: utilize log Gabor bank of filters or Gabor bank of filters that input picture is carried out Filtering Processing, obtain different frequency responses under different scale;
S2. produce characteristic pattern, be specially: calculation procedure S1 obtain each towards the amplitude of the down frequency response of different scale and and local energy, get the ratio of the two, as each towards under characteristic pattern, with difference towards under characteristic pattern integrate and obtain final characteristic pattern;
S3. produce the characteristic pattern after the edge keeps smothing filtering, be specially: utilize the EPS wave filter that the final characteristic pattern that step S2 obtains is carried out Filtering Processing;
S4. produce the contour images after non-classical receptive field suppresses, be specially: the non-classical receptive field wave filter that at first utilizes two-dimentional Gaussian difference function to make up carries out Filtering Processing to the characteristic pattern that obtains among the step S3, obtains the amount of suppression of non-classical receptive field to each pixel; Deduct the amount of suppression of corresponding position then with each grey scale pixel value in the characteristic pattern that obtains among the step S3, the contour images after being inhibited;
S5. binary conversion treatment: the contour images after step S4 is inhibited carries out binary conversion treatment, obtains final contour images.
Beneficial effect of the present invention: profile testing method of the present invention has utilized the phase propetry of image, stability and importance by image phase information, guaranteed the integrality of feature, made profile detect and be not subjected to the influence that brightness of image is inconsistent or contrast changes; And combine the non-classical receptive field characteristic in the human vision mechanism, effectively distinguished objective contour and texture edge, and texture is suppressed.
Description of drawings
Fig. 1 is the schematic flow sheet of a kind of contour of object detection method based on phase propetry of the present invention.
Fig. 2 adopts the inventive method natural image to be carried out each step results figure and the final profile diagram and the nominal contour figure comparison diagram group of actual detected in the example.
Fig. 3 and Fig. 4 adopt the inventive method natural image and the image that added 20dB Gauss noise thereof to be carried out the profile diagram of actual detected and effect contrast figure's group of the profile diagram that detects with isotropic inhibition method in the example.
Embodiment
The present invention is further elaborated below in conjunction with accompanying drawing and concrete example.
Detecting with the profile of a natural image is example, and its image outline that adds behind the 20dB Gauss noise is detected similarly.
Download Basket (basket) image and corresponding nominal contour testing result thereof from the image library website of present internationally recognized checking profile extraction algorithm effect, the image size is 512 * 512, and nominal contour testing result wherein (ground truth) on average obtains after delineating by a plurality of people are manual.The flow process of concrete detection method as shown in Figure 1, detailed process is as follows:
S1. Filtering Processing: the scale parameter number of given log Gabor bank of filters is 4, in 180 ° of scopes, get 8 towards, it is respectively (i-1)/8 towards parameter, (i=1 ... 8), obtain one group of 8 difference towards the log Gabor of 4 different scales (different frequency) wave filter; Fix 4 wave filters then and successively each pixel in the input picture is carried out Filtering Processing, obtain 4 filter responses towards change different scale value structure.
Also can use the Gabor bank of filters to carry out Filtering Processing here.
S2. determine that the most consistent point of phase place is a unique point, produce characteristic pattern: the response at the each point place that filtering obtains is the directive again complex vector of the existing size under 4 different frequencies, to each complex vector ask amplitude and calculate the total amplitude in this some place and, in addition the real part and the addition of imaginary part difference of 4 complex vectors are sued for peace, then to the real part of gained and, imaginary part and squared respectively summation again, the root of finally making even promptly is the local energy at this some place.The characteristic of [0,1] that thereby certain some phase place is consistent asking for exactly of measuring obtains with the total amplitude at this some place with divided by its local energy.Thereby both do ratio and just can try to achieve consistent the measuring of the phase place at this some place and determine whether this point is unique point like this, this normalized can make the characteristic pattern of finally producing not be subjected to the influence that brightness of image is inconsistent or contrast changes, and the feature of extracting is more complete.
Calculated like this one towards under image in phase place consistent the measuring of each pixel, then by to 8 towards the phase place that obtains respectively down consistent measure just to integrate can produce last characteristic pattern.
S3. produce the edge and keep smothing filtering (Edge-Preserving Smoothing filter, EPS) characteristic pattern after: the gradient sensing parameter is 1.6 in the EPS wave filter here, smoothing parameter is 0.8, guarantees to remove texture preserving edge and profile simultaneously effectively after the filtering.
The EPS wave filter that adopts is based on a kind of nonlinear filter of weighted least squares (weighted least squares), by regulating gradient susceptibility parameter and smoothing parameter try to achieve one close as much as possible with original image, while is all level and smooth as much as possible new images in other places except the significant place of original image gradient, this wave filter is adjusted filtering weighting according to the local feature of image like this, thereby reaches the effect of filtering texture and noise preserving edge and profile.
S4. the contour images after producing non-classical receptive field and suppressing: the non-classical receptive field wave filter that utilizes two-dimentional Gaussian difference function to make up carries out Filtering Processing to the characteristic pattern that obtains among the step S3, and the annulus external diameter of non-classical receptive field wave filter is 4 times of internal diameter here.Non-classical receptive field is big or small relevant with the feature difference between them to the inhibition strength of classical receptive field, and difference is big more, and inhibition strength is weak more, otherwise big more, and this inhibition strength is controlled by inhibiting factor, and value is 1.0 here.Result after the Filtering Processing multiply by this inhibiting factor, is the amount of suppression of non-classical receptive field effect, former then characteristic pattern deduct this amount of suppression so that be inhibited after contour images;
Utilize two-dimentional Gaussian difference function to make up the non-classical receptive field wave filter and belong to the state of the art, specifically can be referring to document: Grigorescu C, Petkov N, Westenberg M, Contour detection based on nonclassical receptive field inhibition, IEEE Transactions on Image Processing, vol.12, no.7,729～739,2003, do not remake in this manual and elaborate.
S5. binary conversion treatment: the contour images after step S4 is inhibited carries out binary conversion treatment, obtains final contour images.Here can adopt conventional binarization method to carry out binary conversion treatment, what adopt in the present embodiment is that non-maximum value suppresses and the hysteresis threshold method.
Fig. 2 carries out each step results figure and the final profile diagram and the nominal contour figure comparison diagram group of actual detected for adopting the inventive method to natural image.Wherein: the 2a. original image, the 2b. characteristic pattern, the filtered characteristic pattern of 2c.EPS, the contour images after the 2d. non-classical receptive field suppresses, 2e. nominal contour image, 2f. adopt the inventive method to detect the gained contour images.From figure, can know and find out, when adopting the inventive method to extract the natural image profile, when keeping the integrality of outline line as much as possible, farthest suppress background texture (meadow) and disturb, good to the robustness of nominal contour image.
Fig. 3 and Fig. 4 are respectively and adopt the inventive method natural image to be carried out effect contrast figure's group of the profile diagram that profile diagram reaches and other two kinds of methods detect of actual detected.Wherein: 3a. original image, 3b. adopt the inventive method to detect the gained contour images, and 3c. adopts isotropy to suppress algorithm gained contour images.Remove 4a among Fig. 4 for adding the image behind the 20dB Gauss noise, other figure is the same with Fig. 3.From figure, can know and find out, adopt the inventive method to extract natural image and when adding image outline after making an uproar, the contrast at each some place is incomplete same in the image, basket central part contrast is extremely low and other place is higher relatively, other method of profile that contrast is high relatively also can be extracted, and the effect of extraction the inventive method of central part profile is quite a lot of, the inventive method is in the integrality that keeps outline line as much as possible, farthest suppress background texture (meadow) and noise, its de-noising ability and the robustness of nominal contour image is better than isotropy inhibition method.
For original input image with add the image of making an uproar, the quantitative comparison of method in this enforcement and isotropy inhibition method is as follows: the scale parameter of non-classical receptive field wave filter internal diameter is 2.0 in the isotropy inhibition method, the two profile Performance Evaluation index P values that obtain are respectively 0.36,0.31; The scale parameter of the non-classical receptive field wave filter internal diameter that the method in this enforcement adopts is 1.0, and the two profile Performance Evaluation index P values that obtain are respectively 0.56,0.55.Performance Evaluation index P value is big more to show that the general effect of outstanding profile, compacting background is good more.
Contour of object detection method of the present invention, the advantage of comprehensive utilization phase information and non-classical receptive field rejection characteristic, make profile detect and be not subjected to the influence that brightness of image is inconsistent or contrast changes, effectively raise the ability that from complex scene, detects objective contour.
Those of ordinary skill in the art will appreciate that embodiment described here is in order to help reader understanding's principle of the present invention, should to be understood that protection scope of the present invention is not limited to such special statement and embodiment.Those of ordinary skill in the art can make various other various concrete distortion and combinations that do not break away from essence of the present invention according to these technology enlightenments disclosed by the invention, and these distortion and combination are still in protection scope of the present invention.
Claims (4)
1. the contour of object detection method based on phase propetry is characterized in that, comprises the steps:
S1. Filtering Processing is specially: utilize log Gabor bank of filters or Gabor bank of filters that input picture is carried out Filtering Processing, obtain different frequency responses under different scale;
S2. produce characteristic pattern, be specially: calculation procedure S1 obtain each towards the amplitude of the down frequency response of different scale and and local energy, get the ratio of the two, as each towards under characteristic pattern, with difference towards under characteristic pattern integrate and obtain final characteristic pattern;
S3. produce the characteristic pattern after the edge keeps smothing filtering, be specially: utilize the EPS wave filter that the final characteristic pattern that step S2 obtains is carried out Filtering Processing;
S4. produce the contour images after non-classical receptive field suppresses, be specially: the non-classical receptive field wave filter that at first utilizes two-dimentional Gaussian difference function to make up carries out Filtering Processing to the characteristic pattern that obtains among the step S3, obtains the amount of suppression of non-classical receptive field to each pixel; Deduct the amount of suppression of corresponding position then with each grey scale pixel value in the characteristic pattern that obtains among the step S3, the contour images after being inhibited;
S5. binary conversion treatment: the contour images after step S4 is inhibited carries out binary conversion treatment, obtains final contour images.
2. contour of object detection method according to claim 1 is characterized in that, the gradient sensing parameter is 1.6 in the EPS wave filter described in the step S3, and smoothing parameter is 0.8.
3. contour of object detection method according to claim 1 and 2 is characterized in that, the annulus external diameter of the non-classical receptive field wave filter described in the step S4 is 4 times of internal diameter.
4. contour of object detection method according to claim 3 is characterized in that, what the binary conversion treatment described in the step S5 adopted is that non-maximum value suppresses and the hysteresis threshold method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110163800 CN102222325A (en) | 2011-06-17 | 2011-06-17 | Method for detecting object contour based on phase characteristic |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110163800 CN102222325A (en) | 2011-06-17 | 2011-06-17 | Method for detecting object contour based on phase characteristic |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102222325A true CN102222325A (en) | 2011-10-19 |
Family
ID=44778870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110163800 Pending CN102222325A (en) | 2011-06-17 | 2011-06-17 | Method for detecting object contour based on phase characteristic |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102222325A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102607460A (en) * | 2012-03-13 | 2012-07-25 | 天津工业大学 | Global phase filter method applied to three-dimensional measurement |
| CN103049911A (en) * | 2012-12-20 | 2013-04-17 | 成都理想境界科技有限公司 | Contour detection stability judging method and image searching method |
| CN106373132A (en) * | 2016-08-30 | 2017-02-01 | 刘广海 | Edge detection method based on inhibition internuncial neuron |
| CN109949324A (en) * | 2019-02-01 | 2019-06-28 | 广西科技大学 | Profile testing method based on the non-linear subunit response of non-classical receptive field |
| CN109978898A (en) * | 2019-02-01 | 2019-07-05 | 广西科技大学 | Profile testing method based on vector field energy balane |
| CN113496215A (en) * | 2021-07-07 | 2021-10-12 | 浙江大华技术股份有限公司 | Method and device for detecting human face of living body and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763641A (en) * | 2009-12-29 | 2010-06-30 | 电子科技大学 | Method for detecting contour of image target object by simulated vision mechanism |
| CN101964113A (en) * | 2010-10-02 | 2011-02-02 | 上海交通大学 | Method for detecting moving target in illuminance abrupt variation scene |
| CN102034105A (en) * | 2010-12-16 | 2011-04-27 | 电子科技大学 | Object contour detection method for complex scene |
-
2011
- 2011-06-17 CN CN 201110163800 patent/CN102222325A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763641A (en) * | 2009-12-29 | 2010-06-30 | 电子科技大学 | Method for detecting contour of image target object by simulated vision mechanism |
| CN101964113A (en) * | 2010-10-02 | 2011-02-02 | 上海交通大学 | Method for detecting moving target in illuminance abrupt variation scene |
| CN102034105A (en) * | 2010-12-16 | 2011-04-27 | 电子科技大学 | Object contour detection method for complex scene |
Non-Patent Citations (4)
| Title |
|---|
| 《Neurocomputing》 20110317 Chi Zeng 等 Contour detection based on a non-classical receptive field model with butterfly-shaped inhibition subregions 1527-1534 第74卷, 第10期 * |
| 《中国优秀硕士学位论文全文数据库 信息科技辑 》 20100915 陈霄 基于Gabor变换的视觉特征提取研究 I138-557 , 第 09 期 * |
| 《中国优秀硕士学位论文全文数据库 信息科技辑 》 20110415 曾驰 基于视觉感受野特性的物体轮廓提取算法研究 I138-824 , 第04期 * |
| 《遥感学报》 20070531 肖鹏峰 等 一种基于相位一致的高分辨率遥感图像特征检测方法 303－310 第11卷, 第3期 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102607460A (en) * | 2012-03-13 | 2012-07-25 | 天津工业大学 | Global phase filter method applied to three-dimensional measurement |
| CN102607460B (en) * | 2012-03-13 | 2014-07-30 | 天津工业大学 | Global phase filter method applied to three-dimensional measurement |
| CN103049911A (en) * | 2012-12-20 | 2013-04-17 | 成都理想境界科技有限公司 | Contour detection stability judging method and image searching method |
| CN103049911B (en) * | 2012-12-20 | 2015-07-29 | 成都理想境界科技有限公司 | Contour detecting stability judging method and image search method |
| CN106373132A (en) * | 2016-08-30 | 2017-02-01 | 刘广海 | Edge detection method based on inhibition internuncial neuron |
| CN106373132B (en) * | 2016-08-30 | 2018-10-23 | 广西师范大学 | Edge detection method based on inhibitory interneuron |
| CN109949324A (en) * | 2019-02-01 | 2019-06-28 | 广西科技大学 | Profile testing method based on the non-linear subunit response of non-classical receptive field |
| CN109978898A (en) * | 2019-02-01 | 2019-07-05 | 广西科技大学 | Profile testing method based on vector field energy balane |
| CN109949324B (en) * | 2019-02-01 | 2022-04-22 | 广西科技大学 | Contour detection method based on non-classical receptive field nonlinear subunit response |
| CN109978898B (en) * | 2019-02-01 | 2023-07-18 | 广西科技大学 | Contour detection method based on vector field energy calculation |
| CN113496215A (en) * | 2021-07-07 | 2021-10-12 | 浙江大华技术股份有限公司 | Method and device for detecting human face of living body and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guan et al. | Gaussian scale-space enhanced local contrast measure for small infrared target detection | |
| CN101763641B (en) | Method for detecting contour of image target object by simulated vision mechanism | |
| CN102222325A (en) | Method for detecting object contour based on phase characteristic | |
| CN104200442B (en) | Non-local mean MRI image denoising method based on improved canny rim detections | |
| CN103955922B (en) | Method for detecting flaws of printed fabric based on Gabor filter | |
| CN103530878B (en) | A kind of edge extracting method based on convergence strategy | |
| CN102156996B (en) | Image edge detection method | |
| CN102201120B (en) | Multifeature-based target object contour detection method | |
| CN103440644B (en) | A kind of multi-scale image weak edge detection method based on minimum description length | |
| CN107507209B (en) | Printogram extraction method of polarized SAR image | |
| CN105913382B (en) | The high-fidelity anisotropic filtering method of threshold value optimizing | |
| CN109003233A (en) | A kind of image de-noising method based on adaptive weighting Total Variation | |
| CN105469428B (en) | A kind of detection method of small target based on morphologic filtering and SVD | |
| CN102254304B (en) | Method for detecting contour of target object | |
| CN103208104A (en) | Non-local theory-based image denoising method | |
| CN102034105B (en) | Object contour detection method for complex scene | |
| CN117541654B (en) | Detail enhancement method for high-resolution remote sensing image | |
| CN103971345A (en) | Image denoising method based on improved bilateral filtering | |
| CN109376782B (en) | Support vector machine cataract classification method and device based on eye image features | |
| Ma et al. | A finger vein image quality assessment method using object and human visual system index | |
| CN107134048A (en) | A kind of bill anti-counterfeit discrimination method of Intelligent Recognition watermark feature | |
| CN104573691A (en) | Object contour detection method based on phase characteristics | |
| CN102693530A (en) | Synthetic aperture radar (SAR) image despeckle method based on target extraction and speckle reducing anisotropic diffusion (SRAD) algorithm | |
| Zhang et al. | Region-of-interest extraction based on spectrum saliency analysis and coherence-enhancing diffusion model in remote sensing images | |
| CN110298807A (en) | Based on the domain the NSCT infrared image enhancing method for improving Retinex and quantum flora algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20111019 |