CN105006005A - Evaluation method based on equipartition surface model in color recognition - Google Patents
Evaluation method based on equipartition surface model in color recognition Download PDFInfo
- Publication number
- CN105006005A CN105006005A CN201510498781.6A CN201510498781A CN105006005A CN 105006005 A CN105006005 A CN 105006005A CN 201510498781 A CN201510498781 A CN 201510498781A CN 105006005 A CN105006005 A CN 105006005A
- Authority
- CN
- China
- Prior art keywords
- color
- formula
- equipartition
- color recognition
- facet model
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an evaluation method based on an equipartition surface model in color recognition, which comprises the steps of 1, building an equipartition surface model; 2, optimizing an analytical solution; 3, generating a Guassian based evaluation model; and 4, forming a color evaluation matrix. Construction for a color evaluation matrix function has effects of convenience, rapidness and accuracy for color recognition, particularly for discrimination of similar colors. In actual use, the color evaluation matrix function exceeds Cognex image processing software in color recognition ability, which is particularly reflected in that various colors in color series of white and blue can be distinguished accurately. The evaluation method is already applied to a detection system for 24 kinds of colors of tablets of a company, and can also be applied to the field of multi-color recognition and processing at the same time.
Description
Technical field
The present invention relates to a kind of appraisal procedure of equal facet model, particularly relate to the appraisal procedure based on equal facet model in a kind of colour recognition.
Background technology
Color of object feature can characterize by its color distribution characteristic in HSI color space.For simple application, the threshold range of color of object directly can be provided.But, in actual environment, often due to the difference of the even imaging angle of uneven illumination, make distribution range can not comprise color of object actual spatial distribution under different image-forming condition.Colouring discrimination in the past adopts the methods such as vector, point set distance, one-dimensional analysis, can only distinguish obvious color.And many at color category, and when color is very close, as near-white system and nearly cyan system, classic method just cannot effectively identify, more cannot detect.
Summary of the invention
Technical matters to be solved by this invention is to provide the appraisal procedure based on equal facet model in a kind of colour recognition, it constructs equal facet model, adopt Optimum analyses process, determine three that portray color space point set and divide equally facial index, then completed by valuation functions kernel form the classification of HSI color point set is assessed.
The present invention solves above-mentioned technical matters by following technical proposals: the invention discloses the appraisal procedure based on equal facet model in a kind of colour recognition, it comprises the following steps:
Step one: all facet model builds;
Step 2: Optimum analyses solves;
Step 3: generate the assessment models based on Gaussian;
Step 4: form Color Evaluation matrix.
Preferably, described step one adopts spatial translation, lead-in surface index customized rules.
Preferably, described step 2 introduces Langrange operator.
Preferably, the color data gathering image is transformed into HSI space by described step 4.
Positive progressive effect of the present invention is: the present invention constructs equal facet model, adopts Optimum analyses process, determines three that portray color space point set and divide equally facial index, then is completed by valuation functions kernel form and assess the classification of HSI color point set.The structure of this Color Evaluation matrix function is for colour recognition, particularly the difference of similar color has convenient, fast and effect accurately, in actual use, more than the colour recognition ability of Cognex, be embodied in the multiple color can accurately distinguished in white color system and cyan system especially.This appraisal procedure has been applied in the tablet detection system of 24 kinds of colors of certain company, can also be applied to identification and the process field of many colors simultaneously, simplify the content of evaluates calculation.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the appraisal procedure based on equal facet model in colour recognition of the present invention.
Embodiment
Present pre-ferred embodiments is provided, to describe technical scheme of the present invention in detail below in conjunction with accompanying drawing.
The structure of Color Evaluation matrix function of the present invention is for colour recognition, particularly the difference of similar color has convenient, fast and effect accurately, in actual use, more than the colour recognition ability of Cognex, be embodied in the multiple color can accurately distinguished in white color system and cyan system especially.This appraisal procedure has been applied in the tablet detection system of 24 kinds of colors of certain company, can also be applied to identification and the process field of many colors simultaneously.If discrete builds, many content association can be introduced, but from totally observing, can see that this individual system can be unified into a matrix form, thus simplifying the content of evaluates calculation.
The appraisal procedure based on equal facet model in colour recognition of the present invention comprises the following steps:
Step one, equal facet model builds, and particular content is as follows:
Kernel function adopts the distance form of general closed planar equation structure, and easily and effectively can not solve and divide equally facial index, therefore, the present invention adopts spatial translation, lead-in surface index customized rules, forms good solution procedure by contributing to.
Point set
through spatial translation as shown in the formula the distance of (1) to initial point.
………(1)
In formula
represent point set
central point,
represent respective item average,
represent some lump number.Because in plane to point to the vector of initial point all orthogonal with facial index in arbitrfary point, thus through the plane equation of initial point be as shown in the formula (2):
…………….(2)
In formula
for surface vector; Through the plane of initial point, its facial index can normalizing, as shown in the formula (3):
………………(3)
So the vertical distance form that space point set has been in initial point plane can be reduced to as shown in the formula (4):
…………………..(4)
In formula
represent space point set to the distance of dividing equally face.
Construct equal facet model, can have as shown in the formula (5):
………………..(5)
Here the introducing of constraint condition, denominator situation is independent, and avoid the high order situation introducing parameter, process is simplified, and this thinking has the meaning of expansion in other processes.
Step 2, Optimum analyses solves: introduce Langrange and calculate
, as shown in the formula (6):
…………………(6)
In formula
represent target function value.
As shown in the formula (7) after local derviation:
…………………(7)
Formula (8) is defined as follows in above formula:
……………….(8)
Arrangement has as shown in the formula (9):
………………(9)
Be
the solution procedure of eigenwert and feature value vector.This determinant is 0, has as shown in the formula (10):
……………………(10)
Determine the eigenwert of the square formation of nothing, namely this eigenwert is target function value in fact, gets minimum.Substitute into and have as shown in the formula (11) in matrix:
………………….(11)
Formula (12) is defined as follows in above formula:
…………………….(12)
Substitute in normalizing condition, have as shown in the formula (13):
………………………….(13)
Substitute into former formula, formed as shown in the formula (14):
…………………………..(14)
Substituted in the face equation of initial point, had as shown in the formula (15):
………………(15)
Plane space through initial point is moved to initial central point
, have as shown in the formula (16):
……………….(16)
Get positive sign; Will
definition content substitutes into above formula, has as shown in the formula (17):
…………….(17)
The equation parameter dividing equally face being in initial point is as shown in the formula (18):
……………(18)
In like manner, obtain as shown in the formula (19):
………………(19)
Be verified, pass through.
Note the application difference of the some cloud at initial point place and the some cloud at initial center point place.
In calculating
in amount over overlap time, adopt the some cloud moved to after initial point.In the calculation during heart point, with the some cloud at initial center point place.
Step 3, based on the assessment models of Gaussian, particular content is as follows:
If discrete builds, many content association can be introduced, but from totally observing, can see that this individual system can be unified into a matrix form, thus simplifying the content of evaluates calculation.
First the kernel form of valuation functions is as shown in the formula (20):
……………..(20)
Section 1 be spatial point to the vertical range form of dividing equally face, for as shown in the formula (21):
…………………….(21)
Section 2 in valuation functions e index and Section 3 refer to the description of spatial point perpendicular to the distance of equal separated time.Spatial point needs to be projected on to divide equally on face, and it is as shown in the formula (22) that its form is expressed as matrix pattern:
………………(22)
Integrate and formed as shown in the formula (23):
Order is as shown in the formula (23):
……………..(23)
Then dividing equally the subpoint on face for as shown in the formula (24):
………………(24)
Set up the vertical line distance of straight line very complicated due to direct at three dimensions, general dimensionality reduction, to two-dimensional space, then carries out the description of vertical range, therefore has as shown in the formula (25):
………………(25)
Order is as shown in the formula (26):
,
……….(26)
Step 4, Color Evaluation matrix:
Will
,
with
substitute into valuation functions, as shown in the formula (27):
………………(27)
By gathering the color data of image, being transformed into HSI space, by this evaluating matrix, classification effect clearly can being obtained.When belonging to the HSI data of a certain color, after the process of Color Evaluation matrix function, this assessed value is very large, and close to 1, if offset out this color point set scope a little, then assessed value reduces to 10^ (-10) below immediately.Classification effect is very effective.For colour recognition, can threshold mode be directly adopted to classify.
Above-described specific embodiment; the technical matters of solution of the present invention, technical scheme and beneficial effect are further described; be understood that; the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (4)
1. the appraisal procedure based on equal facet model in colour recognition, it is characterized in that, it comprises:
Step one: all facet model builds;
Step 2: Optimum analyses solves;
Step 3: generate the assessment models based on Gaussian;
Step 4: form Color Evaluation matrix.
2. the appraisal procedure based on equal facet model in colour recognition as claimed in claim 1, it is characterized in that, described step one adopts spatial translation, lead-in surface index customized rules.
3. the appraisal procedure based on equal facet model in colour recognition as claimed in claim 1, it is characterized in that, described step 2 introduces Langrange operator.
4. the appraisal procedure based on equal facet model in colour recognition as claimed in claim 1, it is characterized in that, the color data gathering image is transformed into HSI space by described step 4.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510498781.6A CN105006005A (en) | 2015-08-14 | 2015-08-14 | Evaluation method based on equipartition surface model in color recognition |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510498781.6A CN105006005A (en) | 2015-08-14 | 2015-08-14 | Evaluation method based on equipartition surface model in color recognition |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105006005A true CN105006005A (en) | 2015-10-28 |
Family
ID=54378662
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510498781.6A Pending CN105006005A (en) | 2015-08-14 | 2015-08-14 | Evaluation method based on equipartition surface model in color recognition |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105006005A (en) |
-
2015
- 2015-08-14 CN CN201510498781.6A patent/CN105006005A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10929649B2 (en) | Multi-pose face feature point detection method based on cascade regression | |
Wu et al. | Rachis detection and three-dimensional localization of cut off point for vision-based banana robot | |
JP5721673B2 (en) | Paint color database creation method, creation system, creation program, and recording medium | |
US9805293B2 (en) | Method and apparatus for object recognition in image processing | |
CN103034838B (en) | A kind of special vehicle instrument type identification based on characteristics of image and scaling method | |
Yu et al. | Raisin quality classification using least squares support vector machine (LSSVM) based on combined color and texture features | |
CN106204618A (en) | Product surface of package defects detection based on machine vision and sorting technique | |
CN104966085A (en) | Remote sensing image region-of-interest detection method based on multi-significant-feature fusion | |
CN103617414B (en) | The fire disaster flame of a kind of fire color model based on maximum margin criterion and smog recognition methods | |
CN113128548A (en) | System and method for matching color and appearance of target coating | |
CN105103166A (en) | Systems and methods for texture assessment of a coating formulation | |
EP3828520B1 (en) | Apparatus and method for characterizing effect pigment spots | |
CN102013011A (en) | Front-face-compensation-operator-based multi-pose human face recognition method | |
CN102324032A (en) | Texture feature extraction method for gray level co-occurrence matrix in polar coordinate system | |
US20180225799A1 (en) | System and method for scoring color candidate poses against a color image in a vision system | |
Veras et al. | Discriminability tests for visualization effectiveness and scalability | |
CN111311702B (en) | Image generation and identification module and method based on BlockGAN | |
Ding et al. | FCM image segmentation algorithm based on color space and spatial information | |
CN102855485A (en) | Automatic wheat earing detection method | |
Chowdhury et al. | Vegetables detection from the glossary shop for the blind | |
Kuo et al. | Improving defect inspection quality of deep-learning network in dense beans by using hough circle transform for coffee industry | |
Zhou et al. | Land cover classification from full-waveform lidar data based on support vector machines | |
Wang et al. | Application of hybrid image features for fast and non-invasive classification of raisin | |
CN103870807A (en) | High spectrum mixed nuclear RX anomaly detection method | |
CN104573635B (en) | A kind of little height recognition methods based on three-dimensional reconstruction |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151028 |