CN110489587A - The tire trace image characteristic extracting method of three value mode of Local gradient direction - Google Patents
The tire trace image characteristic extracting method of three value mode of Local gradient direction Download PDFInfo
- Publication number
- CN110489587A CN110489587A CN201910699380.5A CN201910699380A CN110489587A CN 110489587 A CN110489587 A CN 110489587A CN 201910699380 A CN201910699380 A CN 201910699380A CN 110489587 A CN110489587 A CN 110489587A
- Authority
- CN
- China
- Prior art keywords
- image
- value
- gradient direction
- lgdtp
- tire trace
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 10
- 238000003708 edge detection Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 206010039203 Road traffic accident Diseases 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000000052 comparative effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003012 network analysis Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000013517 stratification Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5862—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- 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
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
A kind of tire trace image characteristic extracting method of three value mode of Local gradient direction, by image preprocessing, feature extraction, determines that feature vector, image retrieval step form.The present invention proposes a kind of three value pattern feature of Local gradient direction suitable for tire trace image, gray value is replaced to carry out local grain coding using more stable gradient direction value, threshold value quantizing is carried out to center pixel gradient direction angle, the texture marginal information for generating high quality, improves retrieval accuracy；Feature after feature vector is described carries out similarity calculation with manhatton distance, obtains search result, and retrieval rate is substantially better than other textural characteristics.Have many advantages, such as that tire trace image texture marginal information is clear, retrieval accuracy is high, average precision is high, be suitable for big-sample data, can be used for tire trace image characteristics extraction.
Description
Technical field
The invention belongs to technical field of image processing, and in particular to arrive image characteristic extracting method.
Background technique
In usual traffic accident accident can be explained using the relationship between the live tire impression trace left and material evidence
Process judges both sides' responsibility, and ground tyre trace is often one of the most useful trace evidence, so tire trace retrieval is normal
For public security solve a case or traffic accident treatment in clue obtain.
China's research relevant to tire trace image retrieval is started late, and the research in terms of tire trace more rarely has
Achievement, and there is no the tire trace image measurement database of standard in tire trace image retrieval and sort research field.Mesh
Preceding mainly to carry out systematic searching to tire trace image by the method that traditional characteristic combines, there is no be directed to tire trace image
The characteristics of carry out network analysis.Such as: by analysis SIFT transformation and Gabor wavelet principle, propose based on SIFT-Gabor
The tire trace image steganalysis method of transformation.Feature is extracted in conjunction with unsupervised learning and stratification, one kind is proposed and is based on
The tire trace image search method of rarefaction representation and probability latent semantic analysis.One kind is become based on non-downsampling Contourlet
Change (Non-Subsampled Contourlet Transform, NSCT) and algorithm of co-matrix (Gray Level Co-
Occurrence matrix, GLCM) assemblage characteristic extracting method, and utilize Multistage Support Vector Machine (Support Vector
Machine, SVM) training classifier.This method includes main line by the tire trace image that analysis show that tire and ground are formed
Groove and edge grain essential characteristic.Meanwhile Edge texture line is very complicated, the Edge texture line of tyres of different types has difference
Width and direction, and grain interconnection and winding, trace color is relatively single.
In various texture image descriptors, local binary patterns (Local Binary Patterns, LBP) are a kind of
Popular and powerful image descriptor.Its computation complexity is low, without training study, is easy to Project Realization so calculating
Machine vision and field of image processing have received extensive concern.Based on original LBP method, scholars have also been proposed many improvement
LBP method.Such as: local direction mode (Local Directional Patterns, LDP).The local direction mode of enhancing
(Enhanced Local Directional Patterns, ELDP.Local direction number (Local Directional
Number, LDN) etc..
In technical field of image processing, the technical problem that currently need to urgently solve is to provide a kind of tire trace image
Feature extracting method.
Summary of the invention
Technical problem to be solved by the present invention lies in the above-mentioned prior art is overcome, a kind of tire trace figure is provided
As texture marginal information is clear, retrieval accuracy is high, average precision is high, suitable for the Local gradient direction three of big-sample data
The tire trace image characteristic extracting method of value mode.
Solving technical solution used by above-mentioned technology is to comprise the steps of:
(1) image preprocessing
Every 50~80 progress sizes of class of 30 class of tire trace sample image are chosen from tire trace image data base to return
One changes, gray processing is handled.
(2) feature extraction
1) the image difference G of image in the x-direction is determined with Sobel edge detection methodxWith the image difference G in the direction yy, use
Formula (1) determines the gradient direction angle α (x, y) of each pixel in image
α (x, y)=arctan (Gy/Gx) (1)
2) each gradient direction is determined with three value mode LGDTP method of Local gradient direction in 3 × 3 neighborhood sliding window
Angle value, three value mode LGDTP method of Local gradient direction increase customized threshold value t, giGreater than section [gc-t,gc+ t] } when be 1,
Belonging to this section is 0, and being less than this section is -1, obtains three value encoded radios.
P is the number of neighborhood territory pixel in formula, and R is the radius of neighborhood, 0 < t < 2 π, gcIt is the gradient direction of center pixel
Angle, giIt is the gradient direction angle of its neighborhood territory pixel.
It 3) is positive and negative coding characteristic value by LGDTP Eigenvalues Decomposition, encoded radio is not 1 to be revised as 0, obtains positive coding characteristic
Value LGDTPP, encoded radio is -1 to be revised as 1, remaining encoded radio is revised as 0, obtains negative coding characteristic value LGDTPM。
4) positive coding characteristic value LGDTPPPositive coded image, negative coding characteristic value are constructed instead of the pixel value in image
LGDTPMNegative coded image is constructed instead of the pixel value in image.
(3) feature vector is determined
1) by positive coding characteristic value LGDTPPThe positive coded image and negative coding characteristic value LGDTP constitutedMThe negative coding constituted
Image is respectively equably divided into 3 × 3 sub-blocks, in order from top to bottom, 1~9 subregion of label from left to right, each subregion
Pixel be m × n, m, n be 128 or 256.
2) statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block of positive coded image, cascades all sub-blocks
Histogram；Statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block in negative coded image, cascades all sub-blocks
Histogram cascades the sub-block histogram of positive coded image and negative coded image.
3) using the histogram after cascade as the feature of whole picture tire trace image, positive coded image LGDTPMWith negative coding
Image LGDTPPThe length of feature is respectively 256, and the characteristic length of final coded image LGDTP is 256 × 2 × 5, is expressed as
FLGDTP。
4) by FLGDTPFeature is normalized by (4) formula, the feature vector value after being normalized:
Wherein Fc(t) be feature c t-th of component,It is the feature vector value after normalization.
(4) image retrieval
1) the feature vector value of each width tire trace image is determinedRespectively using every piece image as query image
Measuring similarity is carried out with manhatton distance measure by (5) with each width picture in tire trace image data base:
Wherein d is the distance length between two width tire trace characteristics of image, Xi, XjIndicate the feature of each width tire trace
Vector；
2) use average precision P as retrieval performance evaluation index, (6) determine as the following formula:
Wherein S is comprising correct images number in query result, and K is the total number of images of query result, and K is less than tire trace
The number of every one kind in image data base.
In the step 1) of determination feature vector step (3) of the invention, the m and n be aliquot 3 positive integer,
And it is equal.
Most preferably π/6 t in the step 2) of characteristic extraction step (2) of the invention, in formula (3).
In the step 2) of determination feature vector step (3) of the invention, according to the sky of tire trace image texture information
Between be distributed particularity, the sub-block histogram is the histogram that label is respectively 2,4,5,6,8 subregions.
The present invention is directed to tire trace image data base image and query image, extracts characteristics of image, is existed with query image
Related specy tire trace image is inquired in tire trace image data base, determines tyre kind.By analysis obtain tire with
The tire trace image that ground is formed includes the Edge texture line essential characteristic of main line groove sum.The present invention proposes that one kind is suitable for
The three value pattern feature of Local gradient direction of tire trace image, the more stable gradient direction value of use are carried out instead of gray value
Local grain coding generates the texture marginal information of high quality, mentions by carrying out threshold value quantizing to center pixel gradient direction angle
High retrieval accuracy；Feature after feature vector is described carries out similarity calculation with manhatton distance, obtains search result,
Retrieval rate is substantially better than other textural characteristics.The present invention is clear with tire trace image texture marginal information, retrieval is quasi-
Exactness is high, average precision is high, is suitable for the advantages that big-sample data, can be used for tire trace image characteristics extraction.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention 1.
Fig. 2 is the contrast and experiment figure of 1 method of embodiment Yu 6 kinds of existing texture characteristic extracting methods.
Fig. 3 is the comparative experiments curve of 1 method of embodiment Yu 6 kinds of existing texture characteristic extracting methods.
Specific embodiment
The present invention will be described in further detail with example with reference to the accompanying drawing, but the present invention is not limited to following implementations
Example.
Embodiment 1
The image of the present embodiment tire trace image data base self-built from applicant, including the every 80 width image of class of 30 classes
Total 2400 width, are tested, steps are as follows for the tire trace image characteristic extracting method of three value mode of Local gradient direction
(referring to Fig. 1):
(1) image preprocessing
Every 80 width of class of 30 class of tire trace sample image totally 2400 width is chosen from tire trace image data base, is carried out big
It is small to be normalized to the processing of 384 × 384, gray processing.
(2) feature extraction
1) the image difference G of image in the x-direction is determined with Sobel edge detection methodxWith the image difference G in the direction yy, use
Formula (1) determines the gradient direction angle α (x, y) of each pixel in image
α (x, y)=arctan (Gy/Gx) (7)
2) each gradient direction is determined with three value mode LGDTP method of Local gradient direction in 3 × 3 neighborhood sliding window
Angle value, three value mode LGDTP method of Local gradient direction increase customized threshold value t, giGreater than section [gc-t,gc+ t] } when be 1,
Belonging to this section is 0, and being less than this section is -1, obtains three value encoded radios:
P is the number of neighborhood territory pixel in formula, and R is the radius of neighborhood, and the t of the present embodiment is 2 π, gcIt is the ladder of center pixel
Spend deflection, giIt is the gradient direction angle of its neighborhood territory pixel.
It 3) is positive and negative coding characteristic value by LGDTP Eigenvalues Decomposition, encoded radio is not 1 to be revised as 0, obtains positive coding characteristic
Value LGDTPP, encoded radio is -1 to be revised as 1, remaining encoded radio is revised as 0, obtains negative coding characteristic value LGDTPM。
4) positive coding characteristic value LGDTPPPositive coded image, negative coding characteristic value are constructed instead of the pixel value in image
LGDTPMNegative coded image is constructed instead of the pixel value in image.
(3) feature vector is determined
1) by positive coding characteristic value LGDTPPThe positive coded image and negative coding characteristic value LGDTP constitutedMThe negative coding constituted
Image is respectively equably divided into 3 × 3 sub-blocks, in order from top to bottom, 1~9 subregion of label from left to right, each subregion
Pixel be m × n, m, n 128 of the present embodiment.
2) statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block of positive coded image, cascades all sub-blocks
Histogram；Statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block in negative coded image, cascades all sub-blocks
Histogram cascades the sub-block histogram of positive coded image and negative coded image.According to the sky of tire trace image texture information
Between be distributed particularity, above-mentioned sub-block histogram is the histogram that label is respectively 2,4,5,6,8 subregions.
3) using the histogram after cascade as the feature of whole picture tire trace image, positive coded image LGDTPMWith negative coding
Image LGDTPPThe length of feature is respectively 256, and the characteristic length of final coded image LGDTP is 256 × 2 × 5, is expressed as
FLGDTP。
4) by FLGDTPFeature is normalized by (10) formula, the feature vector value after being normalized:
Wherein Fc(t) be feature c t-th of component,It is the feature vector value after normalization.
(4) image retrieval
1) the feature vector value of each width tire trace image is determinedRespectively using every piece image as query image
Measuring similarity is carried out with manhatton distance measure by (11) with each width picture in tire trace image data base:
Wherein d is the distance length between two width tire trace characteristics of image, Xi, XjIndicate the feature of each width tire trace
Vector.
2) use average precision P as retrieval performance evaluation index, (12) determine as the following formula:
Wherein S is comprising correct images number in query result, and K is the total number of images of query result.
In the present embodiment, using the present embodiment method and local binary patterns LBP method, the local direction mode side LDP
Method, local direction digital modeling LDN method, optimizes the local direction mode side OLDP at enhanced local direction mode ELDP method
Method, three value mode LTP methods of part compare retrieval experiment, part using manhatton distance in tire trace image data base
Binary pattern LBP method, local direction mode LDP method, enhanced local direction mode ELDP method, local direction digital modeling
LDN method, optimization local direction mode OLDP method characteristic dimension are 256 × 5, and three value mode LTP methods of part are by LTP
Eigenvalues Decomposition is positive and negative coding characteristic value, positive coding characteristic value LTPPPositive coded image is constructed instead of the pixel value in image,
Negative coding characteristic value LTPMNegative coded image, positive coding characteristic value LTP are constructed instead of the pixel value in imagePWith negative coding characteristic
Value LTPMAfter cascade composition characteristic dimension be 256 × 5 × 2, according to 2 in step (4)) calculate every kind of being averaged for method look into standard
Rate, experimental result are shown in Table 1 and Fig. 2.The coded image of several method is shown in Fig. 3, be presented in Fig. 3 original image, 6 kinds of methods,
The coded image of the present embodiment, wherein the present embodiment is divided into positive coding characteristic image LGDTPPWith negative coding characteristic image LGDTPM。
The contrast and experiment (K=10) of 1 the present embodiment of table and 6 kinds of methods
| Serial number | Feature | Dimension | Average precision (K=10) |
| 1 | LBP | 1280 | 43.1%, |
| 2 | LDP | 1280 | 45% |
| 3 | ELDP | 1280 | 49.5%, |
| 4 | LDN | 1280 | 47.3% |
| 5 | OLDP | 1280 | 45.6%, |
| 6 | LTP | 2560 | 52.1% |
| 7 | The present embodiment | 2560 | 63.3% |
By table 1, Fig. 2 as it can be seen that when K is 10, the average precision of the present embodiment is 63.3%, the local binary patterns side LBP
It is method, local direction mode LDP method, enhanced local direction mode ELDP method, local direction digital modeling LDN method, optimal
Change local direction mode OLDP method, the average precision of three value mode LTP methods of part is respectively 43.1%, 45%,
49.5%, 47.3%, 45.6%, 52.1%, it can be seen that this method obtains highest average precision.The present embodiment method
Coded image is more clear.
Embodiment 2
The image of the present embodiment tire trace image data base self-built from applicant, including the every 80 width image of class of 30 classes
Total 2400 width, are tested, steps are as follows for the tire trace image characteristic extracting method of three value mode of Local gradient direction:
(1) image preprocessing
The step is same as Example 1.
(2) feature extraction
1) the image difference G of image in the x-direction is determined with Sobel edge detection methodxWith the image difference G in the direction yy, use
Formula (1) determines the gradient direction angle α (x, y) of each pixel in image
α (x, y)=arctan (Gy/Gx) (13)
2) each gradient direction is determined with three value mode LGDTP method of Local gradient direction in 3 × 3 neighborhood sliding window
Angle value, three value mode LGDTP method of Local gradient direction increase customized threshold value t, giGreater than section [gc-t,gc+ t] } when be 1,
Belonging to this section is 0, and being less than this section is -1, obtains three value encoded radios:
P is the number of neighborhood territory pixel in formula, and R is the radius of neighborhood, and the t of the present embodiment is π/18, gcIt is center pixel
Gradient direction angle, giIt is the gradient direction angle of its neighborhood territory pixel.Other steps in the step are same as Example 1.
Other steps are same as Example 1.
Embodiment 3
The image of the present embodiment tire trace image data base self-built from applicant, including the every 80 width image of class of 30 classes
Total 2400 width, are tested, steps are as follows for the tire trace image characteristic extracting method of three value mode of Local gradient direction:
(1) image preprocessing
The step is same as Example 1.
(2) feature extraction
1) the image difference G of image in the x-direction is determined with Sobel edge detection methodxWith the image difference G in the direction yy, use
Formula (1) determines the gradient direction angle α (x, y) of each pixel in image
α (x, y)=arctan (Gy/Gx) (16)
2) each gradient direction is determined with three value mode LGDTP method of Local gradient direction in 3 × 3 neighborhood sliding window
Angle value, three value mode LGDTP method of Local gradient direction increase customized threshold value t, giGreater than section [gc-t,gc+ t] } when be 1,
Belonging to this section is 0, and being less than this section is -1, obtains three value encoded radios:
P is the number of neighborhood territory pixel in formula, and R is the radius of neighborhood, and the t of the present embodiment is 3 pi/2s, gcIt is center pixel
Gradient direction angle, giIt is the gradient direction angle of its neighborhood territory pixel.Other steps in the step are same as Example 1.
Other steps are same as Example 1.
Embodiment 4
In the image preprocessing step (1) of above embodiment 1-3, tire is chosen from tire trace image data base
Totally 2400 width, progress size normalization are the processing of 768 × 768, gray processing to every 80 width of class of 30 class of trace sample image.
The 1 of the determination feature vector step (3) of the present embodiment) step are as follows: by positive coding characteristic value LGDTPPIt constitutes just
Coded image and negative coding characteristic value LGDTPMConstitute negative coded image be respectively equably divided into 3 × 3 sub-blocks, in order from
On down, 1~9 subregion of label from left to right, the pixel of each subregion is m × n, m, n 256 of the present embodiment.The step
Other steps in rapid are same as Example 1.
Other steps are same as Example 1.
Claims (4)
1. a kind of tire trace image characteristic extracting method of three value mode of Local gradient direction, it is characterised in that by following steps
Composition:
(1) image preprocessing
Chosen from tire trace image data base every 50~80 progress size normalizations of class of 30 class of tire trace sample image,
Gray processing processing；
(2) feature extraction
1) the image difference G of image in the x-direction is determined with Sobel edge detection methodxWith the image difference G in the direction yy, use formula
(1) the gradient direction angle α (x, y) of each pixel in image is determined
α (x, y)=arctan (Gy/Gx) (1)
2) each gradient direction angle is determined with three value mode LGDTP method of Local gradient direction in 3 × 3 neighborhood sliding window
Value, three value mode LGDTP method of Local gradient direction increase customized threshold value t, giGreater than section [gc-t,gc+ t] } when be 1, belong to
It is 0 in this section, being less than this section is -1, obtain three value encoded radios:
P is the number of neighborhood territory pixel in formula, and R is the radius of neighborhood, 0 < t < 2 π, gcIt is the gradient direction angle of center pixel, giIt is
The gradient direction angle of its neighborhood territory pixel；
It 3) is positive and negative coding characteristic value by LGDTP Eigenvalues Decomposition, encoded radio is not 1 to be revised as 0, obtains positive coding characteristic value
LGDTPP, encoded radio is -1 to be revised as 1, remaining encoded radio is revised as 0, obtains negative coding characteristic value LGDTPM；
4) positive coding characteristic value LGDTPPPositive coded image, negative coding characteristic value LGDTP are constructed instead of the pixel value in imageMGeneration
Negative coded image is constructed for the pixel value in image；
(3) feature vector is determined
1) by positive coding characteristic value LGDTPPThe positive coded image and negative coding characteristic value LGDTP constitutedMThe negative coded image constituted
3 × 3 sub-blocks are equably respectively divided into, in order from top to bottom, 1~9 subregion of label from left to right, the picture of each subregion
Element is m × n, and m, n are 128 or 256；
2) statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block of positive coded image, cascades all sub-block histograms
Figure；Statistics with histogram is carried out to the gradient direction angle of pixel in each sub-block in negative coded image, cascades all sub-block histograms
Figure cascades the sub-block histogram of positive coded image and negative coded image；
3) using the histogram after cascade as the feature of whole picture tire trace image, positive coded image LGDTPMWith negative coded image
LGDTPPThe length of feature is respectively 256, and the characteristic length of final coded image LGDTP is 256 × 2 × 5, is expressed as FLGDTP；
4) by FLGDTPFeature is normalized by (4) formula, the feature vector value after being normalized:
Wherein Fc(t) be feature c t-th of component,It is the feature vector value after normalization；
(4) image retrieval
1) the feature vector value of each width tire trace image is determinedRespectively using every piece image as query image and wheel
Each width picture in tire mark image database carries out measuring similarity with manhatton distance measure by (5):
Wherein d is the distance length between two width tire trace characteristics of image, Xi, XjIndicate the feature vector of each width tire trace；
2) use average precision P as retrieval performance evaluation index, (6) determine as the following formula:
Wherein S is comprising correct images number in query result, and K is the total number of images of query result, and K is less than tire trace image
The number of every one kind in database.
2. the tire trace image characteristic extracting method of three value mode of Local gradient direction according to claim 1, special
Sign is: in the step 1) for determining feature vector step (3), the m and n are the positive integer of aliquot 3 and equal.
3. the tire trace image characteristic extracting method of three value mode of Local gradient direction according to claim 1, special
Sign is: in the step 2) of characteristic extraction step (2), the t in formula (3) is π/6.
4. the tire trace image characteristic extracting method of three value mode of Local gradient direction according to claim 1, special
Sign is: in the step 2) for determining feature vector step (3), the spatial distribution according to tire trace image texture information is special
Property, it is respectively 2 that the sub-block histogram, which is label, the histogram of 4,5,6,8 subregions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910699380.5A CN110489587B (en) | 2019-07-31 | 2019-07-31 | Tire trace image feature extraction method in local gradient direction three-value mode |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910699380.5A CN110489587B (en) | 2019-07-31 | 2019-07-31 | Tire trace image feature extraction method in local gradient direction three-value mode |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110489587A true CN110489587A (en) | 2019-11-22 |
| CN110489587B CN110489587B (en) | 2023-04-28 |
Family
ID=68548845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910699380.5A Active CN110489587B (en) | 2019-07-31 | 2019-07-31 | Tire trace image feature extraction method in local gradient direction three-value mode |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110489587B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112918956A (en) * | 2021-02-20 | 2021-06-08 | 陆伟凤 | Garbage classification system based on image recognition technology |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100105865A1 (en) * | 2000-10-27 | 2010-04-29 | John Telford | Nucleic acids and proteins from streptococcus groups a & b |
| CN105389573A (en) * | 2015-12-23 | 2016-03-09 | 山东大学 | Face recognition method based on stratified reconstruction in local ternary pattern |
| EP3076367A1 (en) * | 2013-07-22 | 2016-10-05 | Zhejiang University | Method for road detection from one image |
| CN106228163A (en) * | 2016-07-25 | 2016-12-14 | 长安大学 | The local poor ternary sequential image feature that a kind of feature based selects describes method |
| CN106529504A (en) * | 2016-12-02 | 2017-03-22 | 合肥工业大学 | Dual-mode video emotion recognition method with composite spatial-temporal characteristic |
| CN107122396A (en) * | 2017-03-13 | 2017-09-01 | 西北大学 | Three-dimensional model searching algorithm based on depth convolutional neural networks |
| CN109035317A (en) * | 2018-07-04 | 2018-12-18 | 重庆邮电大学 | Illumination reversion and invariable rotary texture expression based on three value mode of gradient local |
| WO2019100436A1 (en) * | 2017-11-22 | 2019-05-31 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face recognition |
| WO2019136894A1 (en) * | 2018-01-10 | 2019-07-18 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face alignment |
-
2019
- 2019-07-31 CN CN201910699380.5A patent/CN110489587B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100105865A1 (en) * | 2000-10-27 | 2010-04-29 | John Telford | Nucleic acids and proteins from streptococcus groups a & b |
| EP3076367A1 (en) * | 2013-07-22 | 2016-10-05 | Zhejiang University | Method for road detection from one image |
| CN105389573A (en) * | 2015-12-23 | 2016-03-09 | 山东大学 | Face recognition method based on stratified reconstruction in local ternary pattern |
| CN106228163A (en) * | 2016-07-25 | 2016-12-14 | 长安大学 | The local poor ternary sequential image feature that a kind of feature based selects describes method |
| CN106529504A (en) * | 2016-12-02 | 2017-03-22 | 合肥工业大学 | Dual-mode video emotion recognition method with composite spatial-temporal characteristic |
| CN107122396A (en) * | 2017-03-13 | 2017-09-01 | 西北大学 | Three-dimensional model searching algorithm based on depth convolutional neural networks |
| WO2019100436A1 (en) * | 2017-11-22 | 2019-05-31 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face recognition |
| WO2019136894A1 (en) * | 2018-01-10 | 2019-07-18 | Zhejiang Dahua Technology Co., Ltd. | Methods and systems for face alignment |
| CN109035317A (en) * | 2018-07-04 | 2018-12-18 | 重庆邮电大学 | Illumination reversion and invariable rotary texture expression based on three value mode of gradient local |
Non-Patent Citations (5)
| Title |
|---|
| FAISAL AHMED: "Automated Facial Expression Recognition Using Gradient-Based Ternary Texture Patterns", 《CHINESE JOURNAL OF ENGINEERING》 * |
| MD AZHER UDDIN: "Dynamic Scene Recognition Using Spatiotemporal Based DLTP on Spark", 《IEEE ACCESS》 * |
| ZEYNAB SHOKOOHI: "Expression recognition using directional gradient local pattern and gradient-based ternary texture patterns", 《2015 2ND INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION AND IMAGE ANALYSIS (IPRIA)》 * |
| 赵飞飞: "基于视频的人脸追踪与识别", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
| 齐美彬: "改进特征与GPU加速的行人检测", 《中国图象图形学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112918956A (en) * | 2021-02-20 | 2021-06-08 | 陆伟凤 | Garbage classification system based on image recognition technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110489587B (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102254303B (en) | Methods for segmenting and searching remote sensing image | |
| CN103810505B (en) | Vehicles identifications method and system based on multiple layer description | |
| CN102096821B (en) | Number plate identification method under strong interference environment on basis of complex network theory | |
| CN107103317A (en) | Fuzzy license plate image recognition algorithm based on image co-registration and blind deconvolution | |
| Wang et al. | License plate segmentation and recognition of Chinese vehicle based on BPNN | |
| CN103473551A (en) | Station logo recognition method and system based on SIFT operators | |
| CN104182763A (en) | Plant type identification system based on flower characteristics | |
| CN101777125B (en) | Method for supervising and classifying complex category of high-resolution remote sensing image | |
| CN105069481A (en) | Multi-label natural scene classification method based on spatial pyramid and sparse coding | |
| CN104200228A (en) | Recognizing method and system for safety belt | |
| CN102147812A (en) | Three-dimensional point cloud model-based landmark building image classifying method | |
| CN103955952A (en) | Extraction and description method for garment image color features | |
| CN107085731A (en) | A kind of image classification method based on RGB D fusion features and sparse coding | |
| CN105139011A (en) | Method and apparatus for identifying vehicle based on identification marker image | |
| CN112800980A (en) | SAR target recognition method based on multi-level features | |
| CN116704490B (en) | License plate recognition method, license plate recognition device and computer equipment | |
| Elsayad et al. | A new spatial weighting scheme for bag-of-visual-words | |
| Amores et al. | Boosting contextual information in content-based image retrieval | |
| CN106844785A (en) | Saliency segmentation-based content-based image retrieval method | |
| CN105930497A (en) | Image edge and line feature based three-dimensional model retrieval method | |
| CN105654122A (en) | Spatial pyramid object identification method based on kernel function matching | |
| CN110309793A (en) | A kind of SAR target identification method based on video bits layering interpretation | |
| CN110489587A (en) | The tire trace image characteristic extracting method of three value mode of Local gradient direction | |
| CN108805183A (en) | A kind of image classification method of fusion partial polymerization descriptor and local uniform enconding | |
| CN113516123A (en) | Detection and identification method for tire embossed characters |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |