CN112488958A - Image contrast enhancement method based on scale space - Google Patents
Image contrast enhancement method based on scale space Download PDFInfo
- Publication number
- CN112488958A CN112488958A CN202011477203.1A CN202011477203A CN112488958A CN 112488958 A CN112488958 A CN 112488958A CN 202011477203 A CN202011477203 A CN 202011477203A CN 112488958 A CN112488958 A CN 112488958A
- Authority
- CN
- China
- Prior art keywords
- image
- pyramid
- gaussian
- scale
- feature
- 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
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000004364 calculation method Methods 0.000 claims abstract description 24
- 238000001514 detection method Methods 0.000 claims abstract description 15
- 239000011229 interlayer Substances 0.000 claims abstract description 9
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 239000010410 layer Substances 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 4
- 230000004044 response Effects 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 12
- 230000002401 inhibitory effect Effects 0.000 abstract 1
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 102100037364 Craniofacial development protein 1 Human genes 0.000 description 2
- 101000880187 Homo sapiens Craniofacial development protein 1 Proteins 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an image contrast enhancement method based on scale space, which comprises the following steps: s1, preprocessing the original image data to obtain input image data; s2, constructing an image pyramid for the input image data and an image difference pyramid constructed by the image pyramid; s3, overlapping the interlayer images in the same group of the image difference pyramid obtained in the S2 to obtain a group of new image scale spaces; s4, and carrying out feature detection, feature description and feature matching in a new image scale space of the image difference pyramid constructed in S2 and the S3 building block. The invention enhances the image contrast by overlapping and fusing the images under different scales in the scale space, enhances the image contrast on the premise of not changing the characteristic structure of the image, and simultaneously has a certain inhibiting effect on the image noise. Meanwhile, the method has great advantages in calculation amount, and good effects are achieved in the traditional image recognition, image matching and image splicing algorithms.
Description
Technical Field
The invention relates to the technical field of image processing and the like, relates to an image contrast enhancement technology, and particularly relates to an image contrast enhancement method based on a scale space.
Background
With the development of science and technology, the artificial intelligence industry rises rapidly, and the field of computer vision is more colorful. Various algorithms are provided, so that the landing of various artificial intelligence products, such as fingerprint identification, face identification, license plate identification and the like, is realized, and great convenience is brought to human life. The scientific technology continuously provides innovation and solves the problems before various problems.
In the image processing direction, various algorithms have been proposed in academic circles such as image denoising and image contrast enhancement, and great potential is brought into play under image noises with different properties. The quality of an image is also a big factor influencing practical application effects such as image recognition, image matching, image splicing and the like, and various existing filtering algorithms can smooth off some detailed information of the image while filtering abnormal information, so that under a low-quality image environment, an image processing related algorithm has an unobvious effect on contrast enhancement of a low-quality image, and actual requirements are difficult to meet.
Disclosure of Invention
Aiming at the problem that the contrast enhancement effect of the low-quality image is not obvious by the existing image processing correlation algorithm, the invention aims to provide an image contrast enhancement method based on a scale space so as to solve the problem of the low-quality image enhancement correlation and improve the effects in the applications of image identification, image matching and image splicing.
The invention adopts the following technical scheme:
an image contrast enhancement method based on scale space comprises the following steps:
s1, preprocessing the original image data to obtain input image data;
s2, constructing an image pyramid for the input image data and an image difference pyramid constructed by the image pyramid;
s3, overlapping the interlayer images in the same group of the image difference pyramid obtained in the S2 to obtain a group of new image scale spaces;
s4, feature detection, feature description and feature matching in the new image scale space of the S3 building block.
Preferably, in S1, the original image data is subjected to denoising processing to obtain input image data.
Preferably, when denoising the original image data, the filter is selected according to the noise property of the image.
Preferably, in S2, the image pyramid is an image gaussian difference pyramid, and the image difference pyramid is an image gaussian difference pyramid.
Preferably, the gaussian functions with different scale factors are used for performing convolution calculation on the input image data to obtain an image gaussian pyramid, and a calculation formula of the image gaussian pyramid is as follows:
L(x,y,σ)＝G(x,y,σ)*I(x,y)
wherein, denotes convolution operation, σ is a scale factor, G (x, y, σ) is a gaussian function with the scale factor, I (x, y) is input image data, L (x, y, σ) is a gaussian pyramid layer image obtained after convolution, x is a horizontal scale of the image, y is a vertical coordinate of the image, m is a gaussian kernel length, and n is a gaussian kernel width.
Preferably, in S2, images of two adjacent gaussian scales in the same group of the image gaussian pyramid are subtracted to obtain an image gaussian difference pyramid, where a calculation formula of the image gaussian difference pyramid is as follows:
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y)
＝L(x,y,kσ)-L(x,y,σ)
wherein D (x, y, σ) is a gaussian difference pyramid of the image, G (x, y, k σ) is a gaussian function with a scale factor k σ, G (x, y, σ) is a gaussian function with a scale factor attached thereto, denotes a convolution operation, I (x, y) is input image data, L (x, y, k σ) is a gaussian pyramid layer image with a scale k σ, and L (x, y, σ) is a gaussian pyramid layer image with a scale σ.
Preferably, in S3, different energy coefficients are superimposed on the interlayer image in the same group of the image gaussian difference pyramid, so as to obtain a new image scale space, where a calculation formula of the new image scale space is as follows:
Di(x,y)＝a*Di(x,y)+b*Di+2(x,y)
wherein a and b are energy coefficients and range is (0.0, 1.0); i is the index of the image of the Gaussian difference pyramid layer, represents the convolution operation, Di(x, y) is a new image scale space.
Preferably, in S4, the feature detection process includes:
and comparing the current pixel point with a plurality of adjacent pixel points with the same scale and a plurality of pixel point response values corresponding to the upper and lower adjacent scales, judging whether the current pixel point is an extreme point, if so, finding the accurate position of the point on the sub-pixel in a curve fitting mode, recording the accurate position as a characteristic point, and finishing the characteristic detection.
Preferably, in S4, the feature description describes the feature points according to the SIFT operator, including the calculation of the feature point principal direction and the descriptor.
Preferably, in S4, the feature matching process includes: and carrying out primary coarse matching through the Euclidean distance between the feature points, and then carrying out secondary accurate matching by using a random sampling consistency algorithm to obtain an image feature matching result.
The invention has the following beneficial effects:
the image contrast enhancement method based on the scale space enhances the image contrast by overlapping and fusing images under different scales in the scale space, enhances the image contrast on the premise of not changing the characteristic structure of the image, and simultaneously has a certain inhibition effect on image noise. In addition, the method of the invention also has great advantages in the calculation amount, and has good effects in the traditional algorithms of image recognition, image matching and image splicing.
Further, in S3, different energy coefficients are superimposed on the interlayer images in the same group of the image gaussian difference pyramid to obtain a new image scale space, and this method depends on the characteristic that the images in the scale space are features of the same input image at different scales, so that the operation does not cause abnormal changes in the image structure.
Drawings
FIG. 1 is a flow chart of a method for scale-space based image contrast enhancement according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the pyramid of the image described in step S2 according to an embodiment of the present invention;
fig. 3 is a schematic diagram of the image differential pyramid calculation described in step S2 according to the embodiment of the present invention;
fig. 4 is a comparison of the original scale space and the completely new scale space after the superposition of the images with different scales described in step S3 provided by the embodiment of the present invention.
Fig. 5 is a comparison of the effects of feature detection, description, and matching described in step S4 provided in the example of the invention.
Detailed Description
The present invention will now be described in further detail with reference to the attached drawings, which are illustrative, but not limiting, of the present invention.
Referring to fig. 1-4, the image contrast enhancement method based on scale space of the present invention includes the following steps:
s1, preprocessing the original image data to obtain input image data, wherein the preprocessing comprises image denoising;
s2, constructing an image scale space for the input image, wherein the image scale space comprises an image pyramid and an image difference pyramid, and the image scale space adopts a Gaussian pyramid and a Gaussian difference pyramid.
And S3, overlapping the interlayer images in the same group of the image differential pyramid to obtain the scale space of the method.
And S4, carrying out feature detection, feature description and feature matching in the Scale space of the invention, and carrying out related experiments by using a Scale-invariant feature transform (SIFT) operator.
S5, the index for evaluating the method is the comparison between the image scale space visual effect and the feature matching result.
In step S1, preprocessing such as denoising is performed on the image data, and an average filter or other filters are used to select the image data according to the noise property of the image, so as to obtain input image data.
In the image scale space described in step S2, the gaussian function with different scale factors is used to perform convolution calculation on the input image to obtain a gaussian pyramid of the image. The calculation formula is as follows:
L(x,y,σ)＝G(x,y,σ)*I(x,y)
where σ is a scale factor, G (x, y, σ) is a gaussian function with the scale factor attached, I (x, y) is an input image, and L (x, y, σ) is a gaussian pyramid layer image obtained after convolution.
Step S2 further includes constructing an image difference pyramid, and subtracting two adjacent gaussian-scale images in the same group of gaussian pyramids to obtain a gaussian difference pyramid layer image. The calculation formula is as follows:
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y)
＝L(x,y,kσ)-L(x,y,kσ)
wherein D (x, y, σ) is a gaussian difference pyramid layer image obtained by subtraction.
In step S3, the interlayer images are superimposed with different energy coefficients in the same group of the gaussian difference image pyramid to enhance the image contrast. The method is based on the characteristic that the image in the scale space is the same input image with the characteristics of different scales, so the operation does not cause the abnormal change of the image structure. The calculation formula is as follows:
Di(x,y)＝a*Di(x,y)+b*Di+2(x,y)
wherein a and b are energy coefficients in the range of (0.0, 1.0); i is a Gaussian difference pyramid level image index, Di(x, y) is the enhanced image.
In step S4, comparison tests such as feature detection, feature description, feature matching, and the like are performed in the original Scale space and the new Scale space, and a Scale-invariant feature transform (SIFT) operator is taken as an example. The characteristic detection is calculated in a Gaussian difference pyramid, and the current pixel point is compared with 8 adjacent points with the same scale and 9 x 2 points corresponding to upper and lower adjacent scales to count up to 26 point response values, so that whether the current pixel point is an extreme point is judged. If yes, the accurate position of the sub-pixel is found through a curve fitting mode, and the position is recorded as a feature point.
In the step S4, the feature description is performed on the feature points according to SIFT operators; in the characteristic matching stage, firstly, primary coarse matching is carried out through the Euclidean distance between characteristic points, and then secondary accurate matching is carried out by using the RANSAC algorithm. Further, please refer to the objects or advantages of the supplementary explanation arrangement according to the claims.
Step S5 is to use the visual effect of the image in two different scales and space and the result of feature matching as the evaluation index of the method.
Examples
Referring to fig. 1, a flow chart of an image contrast enhancement method based on scale space includes the following steps:
and S1, carrying out denoising preprocessing on the original image data to obtain input image data.
And S2, constructing an image pyramid and an image difference pyramid.
And S3, overlapping and fusing the images under different scales in the same group of the image difference pyramid by using different energy coefficients.
And S4, performing feature detection, description and matching in two different scale spaces, and outputting the result.
Referring to fig. 2, the image pyramid is constructed in the following calculation manner:
the scale space L (x, y, σ) is defined as the convolution of the original image I (x, y) with a two-dimensional gaussian function of variable scale, and has the formula L (x, y, σ) ═ G (x, y, σ) × I (x, y), where x represents the convolution operation,
the image scale of each layer in the Gaussian pyramid group is sigmai＝kσ0Wherein
S is the number of layers used for feature point detection, N is the number of layers in each group of Gaussian pyramid, sigma0Is the scale of the input image.
Images within the same group are convolved by a Gaussian convolution factor of
And convolving the input image.
Referring to fig. 3, the construction step of the image differential pyramid includes the following steps:
the image difference pyramid image is obtained by subtracting two adjacent layers of the pyramid, and the calculation formula is as follows:
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y)
＝L(x,y,kσ)-L(x,y,kσ)
referring to fig. 4, the specific method for superimposing images of different scales of the image difference pyramid is as follows:
the superposition of different energy coefficients is carried out on the interlayer images in the same group, so that the characteristic structure of the images cannot be abnormally changed, because the images in the same group are obtained by carrying out convolution on the reference layer images, the blurring degree is different, and the image sizes are consistent. The calculation formula is as follows:
Di(x,y)＝a*Di(x,y)+b*Di+2(x,y)
referring to fig. 5, feature detection, description, and matching are performed in a scale space:
the characteristic detection is obtained by calculation in three continuous layers of a scale space, and the current pixel point is compared with 8 adjacent points with the same scale and 9-2 points corresponding to the upper and lower adjacent scales, and the total response value of 26 points is judged to determine whether the current pixel point is an extreme point. If yes, the accurate position of the sub-pixel is found through a curve fitting mode, and the position is recorded as a feature point.
The feature description needs to calculate information such as a main direction, a descriptor and the like for a feature point, and the information is calculated by using a calculation method provided by a Scale-invariant feature transform (SIFT) operator privately.
Calculating Euclidean distance between feature points extracted from quasi-matching image if Dismin<th*Dissec_min,0.0<th<1.0
Then Dis is recordedminThe pair of matching point pairs. And reversely calculating the characteristic points of the two images, and recording the point pair as a high-quality point pair if the matching point pair also meets the calculation formula. And purifying all the matching point pairs by using an RANSAC algorithm to obtain a final matching result.
In this embodiment, the feature matching point logarithm (CNT), the matching optimal point logarithm (BCNT) index, the feature matching result extracted from the original scale space and the scale space calculated by the method provided by the present invention, and the visual effect of the image are evaluated, as shown in table 1.
| Methods | CNT | BCNT |
| Original method | 107 | 48 |
| The method of the invention | 172 | 79 |
TABLE 1
As can be seen from table 1, the method proposed in the present invention has a very significant effect in two indexes.
The invention discloses an image contrast enhancement method based on a scale space, which is used for carrying out superposition fusion on images under different scales by using different energy coefficients. According to the method, the image characteristics in the scale space are used, the effect generated by superposition not only enhances the image contrast, but also has an obvious effect on image denoising, and the quality of the detected new features is higher. In addition, the method of the invention also has strong advantages in the aspect of calculation amount, and the performance and the effect of the traditional algorithm related to image processing are greatly improved.
The above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.
Claims (10)
1. An image contrast enhancement method based on scale space is characterized by comprising the following steps:
s1, preprocessing the original image data to obtain input image data;
s2, constructing an image pyramid for the input image data and an image difference pyramid constructed by the image pyramid;
s3, overlapping the interlayer images in the same group of the image difference pyramid obtained in the S2 to obtain a group of new image scale spaces;
s4, feature detection, feature description and feature matching in the new image scale space of the S3 building block.
2. The method of claim 1, wherein in S1, denoising processing is performed on original image data to obtain input image data.
3. The method of claim 2, wherein the denoising of the original image data is performed by selecting a filter according to the noise property of the image.
4. The method of claim 1, wherein in step S2, the image pyramid is an image gaussian pyramid, and the image difference pyramid is an image gaussian difference pyramid.
5. The method of claim 4, wherein the Gaussian functions with different scale factors are used to perform convolution calculation on the input image data to obtain the Gaussian pyramid of the image, and the calculation formula of the Gaussian pyramid of the image is as follows:
L(x,y,σ)＝G(x,y,σ)*I(x,y)
wherein, denotes convolution operation, σ is a scale factor, G (x, y, σ) is a gaussian function with the scale factor, I (x, y) is input image data, L (x, y, σ) is a gaussian pyramid layer image obtained after convolution, x is a horizontal scale of the image, y is a vertical coordinate of the image, m is a gaussian kernel length, and n is a gaussian kernel width.
6. The method of claim 4, wherein in step S2, the subtraction is performed between two adjacent gaussian-scale images in the same group of gaussian pyramids of the image to obtain a gaussian difference pyramid of the image, and the calculation formula of the gaussian difference pyramid of the image is as follows:
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y)
＝L(x,y,kσ)-L(x,y,σ)
wherein D (x, y, σ) is a gaussian difference pyramid of the image, G (x, y, k σ) is a gaussian function with a scale factor k σ, G (x, y, σ) is a gaussian function with a scale factor attached thereto, denotes a convolution operation, I (x, y) is input image data, L (x, y, k σ) is a gaussian pyramid layer image with a scale k σ, and L (x, y, σ) is a gaussian pyramid layer image with a scale σ.
7. The method of claim 4, wherein in step S3, different energy coefficients are superimposed on the interlayer images in the same group of Gaussian difference pyramids of the images to obtain a new image scale space, and a calculation formula of the new image scale space is as follows:
Di(x,y)＝a*Di(x,y)+b*Di+2(x,y)
wherein a and b are energy coefficients and range is (0.0, 1.0); i is the index of the image of the Gaussian difference pyramid layer, represents the convolution operation, Di(x, y) is a new image scale space.
8. The method of claim 4, wherein in step S4, the feature detection process includes:
and comparing the current pixel point with a plurality of adjacent pixel points with the same scale and a plurality of pixel point response values corresponding to the upper and lower adjacent scales, judging whether the current pixel point is an extreme point, if so, finding the accurate position of the point on the sub-pixel in a curve fitting mode, recording the accurate position as a characteristic point, and finishing the characteristic detection.
9. The method of claim 4, wherein in the step S4, the description of the feature points by the feature description according to SIFT operator includes the calculation of feature point principal direction and descriptor.
10. The method of claim 4, wherein in step S4, the feature matching process includes: and carrying out primary coarse matching through the Euclidean distance between the feature points, and then carrying out secondary accurate matching by using a random sampling consistency algorithm to obtain an image feature matching result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011477203.1A CN112488958A (en) | 2020-12-15 | 2020-12-15 | Image contrast enhancement method based on scale space |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011477203.1A CN112488958A (en) | 2020-12-15 | 2020-12-15 | Image contrast enhancement method based on scale space |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112488958A true CN112488958A (en) | 2021-03-12 |
Family
ID=74917084
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011477203.1A Pending CN112488958A (en) | 2020-12-15 | 2020-12-15 | Image contrast enhancement method based on scale space |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112488958A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113255523A (en) * | 2021-05-26 | 2021-08-13 | 上海弘遥电子研究开发有限公司 | Method, system, device and storage medium for improving gesture recognition precision and recognition |
| CN114170445A (en) * | 2022-02-10 | 2022-03-11 | 河北工业大学 | Indoor smoke environment image matching method suitable for fire-fighting robot |
| CN116385280A (en) * | 2023-01-09 | 2023-07-04 | 爱芯元智半导体(上海)有限公司 | Image noise reduction system and method and noise reduction neural network training method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103593832A (en) * | 2013-09-25 | 2014-02-19 | 重庆邮电大学 | Method for image mosaic based on feature detection operator of second order difference of Gaussian |
| CN104834931A (en) * | 2015-03-13 | 2015-08-12 | 江南大学 | Improved SIFT algorithm based on wavelet transformation |
| CN108304883A (en) * | 2018-02-12 | 2018-07-20 | 西安电子科技大学 | Based on the SAR image matching process for improving SIFT |
| CN108830237A (en) * | 2018-06-21 | 2018-11-16 | 北京师范大学 | A kind of recognition methods of human face expression |
| CN109785371A (en) * | 2018-12-19 | 2019-05-21 | 昆明理工大学 | A kind of sun image method for registering based on normalized crosscorrelation and SIFT |
| CN111223068A (en) * | 2019-11-12 | 2020-06-02 | 西安建筑科技大学 | Retinex-based self-adaptive non-uniform low-illumination image enhancement method |
-
2020
- 2020-12-15 CN CN202011477203.1A patent/CN112488958A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103593832A (en) * | 2013-09-25 | 2014-02-19 | 重庆邮电大学 | Method for image mosaic based on feature detection operator of second order difference of Gaussian |
| CN104834931A (en) * | 2015-03-13 | 2015-08-12 | 江南大学 | Improved SIFT algorithm based on wavelet transformation |
| CN108304883A (en) * | 2018-02-12 | 2018-07-20 | 西安电子科技大学 | Based on the SAR image matching process for improving SIFT |
| CN108830237A (en) * | 2018-06-21 | 2018-11-16 | 北京师范大学 | A kind of recognition methods of human face expression |
| CN109785371A (en) * | 2018-12-19 | 2019-05-21 | 昆明理工大学 | A kind of sun image method for registering based on normalized crosscorrelation and SIFT |
| CN111223068A (en) * | 2019-11-12 | 2020-06-02 | 西安建筑科技大学 | Retinex-based self-adaptive non-uniform low-illumination image enhancement method |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113255523A (en) * | 2021-05-26 | 2021-08-13 | 上海弘遥电子研究开发有限公司 | Method, system, device and storage medium for improving gesture recognition precision and recognition |
| CN113255523B (en) * | 2021-05-26 | 2022-07-12 | 上海弘遥电子研究开发有限公司 | Method, system, device and storage medium for improving gesture recognition precision and recognition |
| CN114170445A (en) * | 2022-02-10 | 2022-03-11 | 河北工业大学 | Indoor smoke environment image matching method suitable for fire-fighting robot |
| CN114170445B (en) * | 2022-02-10 | 2022-04-12 | 河北工业大学 | Indoor smoke environment image matching method suitable for fire-fighting robot |
| CN116385280A (en) * | 2023-01-09 | 2023-07-04 | 爱芯元智半导体(上海)有限公司 | Image noise reduction system and method and noise reduction neural network training method |
| CN116385280B (en) * | 2023-01-09 | 2024-01-23 | 爱芯元智半导体(上海)有限公司 | Image noise reduction system and method and noise reduction neural network training method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112488958A (en) | Image contrast enhancement method based on scale space | |
| Park et al. | Double JPEG detection in mixed JPEG quality factors using deep convolutional neural network | |
| CN106355577B (en) | Rapid image matching method and system based on significant condition and global coherency | |
| CN115294158A (en) | Hot continuous rolling strip steel image segmentation method based on machine vision | |
| CN112819772B (en) | High-precision rapid pattern detection and recognition method | |
| CN108416789A (en) | Method for detecting image edge and system | |
| CN112017223B (en) | Heterologous image registration method based on improved SIFT-Delaunay | |
| Yu et al. | A new edge detection approach based on image context analysis | |
| CN116228747B (en) | Metal cabinet processing quality monitoring method | |
| CN110782477A (en) | Moving target rapid detection method based on sequence image and computer vision system | |
| CN112580661B (en) | Multi-scale edge detection method under deep supervision | |
| CN108550166B (en) | Spatial target image matching method | |
| CN109509163B (en) | FGF-based multi-focus image fusion method and system | |
| CN113870235A (en) | Method for detecting defects of circular stamping part based on quantum firework arc edge extraction | |
| Cao et al. | Infrared small target detection based on derivative dissimilarity measure | |
| CN111915486B (en) | Confrontation sample defense method based on image super-resolution reconstruction | |
| CN110648316A (en) | Steel coil end face edge detection algorithm based on deep learning | |
| CN108205657A (en) | Method, storage medium and the mobile terminal of video lens segmentation | |
| CN106600613A (en) | Embedded GPU-based improved LBP infrared target detection method | |
| CN111223063A (en) | Finger vein image NLM denoising method based on texture features and binuclear function | |
| CN111950635B (en) | Robust feature learning method based on layered feature alignment | |
| CN108596928A (en) | Based on the noise image edge detection method for improving Gauss-Laplace operator | |
| CN108470345A (en) | A kind of method for detecting image edge of adaptive threshold | |
| CN111368856A (en) | Spine extraction method and device of book checking system based on vision | |
| CN108010076B (en) | End face appearance modeling method for intensive industrial bar image detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210312 |
|
| RJ01 | Rejection of invention patent application after publication |