CN1821750A - Digital photoelectric grease integration color detecting system and its detecting method - Google Patents
Digital photoelectric grease integration color detecting system and its detecting method Download PDFInfo
- Publication number
- CN1821750A CN1821750A CN 200610013384 CN200610013384A CN1821750A CN 1821750 A CN1821750 A CN 1821750A CN 200610013384 CN200610013384 CN 200610013384 CN 200610013384 A CN200610013384 A CN 200610013384A CN 1821750 A CN1821750 A CN 1821750A
- Authority
- CN
- China
- Prior art keywords
- color
- sample
- grease
- lovibond
- light
- 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
Images
Landscapes
- Spectrometry And Color Measurement (AREA)
Abstract
This invention relates to a photoelectric integral color test system including a light source and a NN processor characterizing that the light of the source is reflected by a diffusing reflection board to generate an uniform illumination field on the chip channel, the beam of which is emitted to the reflection mirror then enters into a charge coupling detector, a data collecting card and the NN processor. A test method for digitalized photoelectric grease integration color test system includes the following steps: 1, setting up a Lovibond color comparison chip database, 2, setting up a grease sample database, 3, a NN test stage and 4, test stage, which can search for the Lovibond expression of grease in the database to realize the automatic evaluation of the sample.
Description
Technical field
The present invention relates to a kind of photoelectric integration chromatic measuring system, particularly relate to the instability compensation of light source in the color and luster analytic process and the calibration problem of database, belong to photoelectric detection technology field.
Background technology
Color and luster is one of important quality index of grease, the outward appearance foundation that simultaneously also is grease deciding grade and level, price and consumer when shelf are chosen, thereby become the conventional project of oil quality check.The factor that influences grease color has many aspects, and therefore most importantly processing technology and condition, measure grease color to improving grease processing technique and condition, and the quality that makes the grease of high-quality and evaluation grease is significant.
The assay method of grease color commonly used has visual method, potassium dichromate method and Lovibond colourimetry.Visual method is by human eye Direct observation grease color, and potassium dichromate method is that oil sample and the potassium dichromate sulfuric acid standard solution for preparing are carried out colorimetric.The Lovibond colourimetry is to compare with the synthetic color and the oil sample color of Lovibond standard than color chips.There is following problem in the Lovibond colourimetry: for same sample, can obtain Different Results because of different operating person's vision difference, promptly have inevitably artificial deviation; Can only carry out intermittent operation, can not realize serialization and robotization, labour intensity is big; Simultaneously, Lovibond can be along with the prolongation of service time " misalignment " than the color of color chips.
The automatic colorimetric method of realizing the grease color analysis mainly contains three kinds: (1) is based on the automatic colorimetric system of spectrophotometric method, because cost is higher, and the inreal realization commercialization of such tintmeter; (2) based on the automatic colorimetric method of many optical filters, mainly utilize many optical filters to extract the principal character point of the spectrum of color correspondence, based on the auto colorimeter commercialization of this method, shortcoming is that the optical characteristics of optical filter is had relatively high expectations, and the cost of instrument is also higher; (3) colored CCD and scanner are united the method for use, in fact strict, the method is not automatic colorimetric system, a kind of full-automatic edible oil colorimeter is invented by the U.S. McCloskey scientific instrument company that Texas Instrument adopts, its measurement range is European color standard yellow 40/ red 22 and U.S. AOCS color standard yellow 40/ red 22, because its measurement range is limited, therefore be not widely used.Britain Tintometer company has released the AF960 auto colorimeter in 1984, this instrument can be measured European color standard yellow 20/ red 4 can measure U.S. AOCS color standard yellow 70/ red 20 again, but discover that AF960 and visual Lovibond tintmeter AF710 can not form good correlativity, therefore be not widely used.1996,12 national joint researches such as the U.S., Canada were adopted and are strengthened red value and the yellow value that online tintmeter PFX990 measures grease color.The result shows that red value and Lovibond have correlativity preferably.
Summary of the invention
The invention provides the technical scheme of a kind of digital photoelectric grease integration color detecting system and detection method thereof, to solve the technical matters that exists in the prior art: system aspects, utilize electrooptical device to replace human eye to receive the color that Lovibond forms than color chips, and color is represented with the RGB color space; Theoretical side, the error at the instability of light-source brightness causes proposes a kind of light path reference entry and proofreaies and correct.Utilize the standard Lovibond to set up Lovibond than color chips standard database than the various combined colors of color chips.After recording the oil sample color, data are carried out match, in standard database, retrieve then with the Neural Network Based Nonlinear approximating method.
The structural representation of native system comprises three major parts altogether shown in figure one: be light path part, photoelectricity collecting part, ANN (Artificial Neural Network) Control part.
(1) light path
Usually design an optical system, at first want the light source of taking into account system.The color of object is different with the difference of lighting source, the throw light on object of same color of Different Light, and people can observe different colors, record different RGB.Usually require the approximate CIE standard light source of recommending of spectral power distribution of light source, require the spectral transmittance of optical medium in visible-range should be similar to constant simultaneously.According to the light path of above this chromatic measuring system of principle shown in figure one.The light that light source sends passes through the catoptron of the beam channel in-position 4 of position 3 after the diffuse reflector diffuse reflection of position 2, be imaged on then on the CCD light-sensitive surface of position 5.When setting up database, the position 8 in the position 3 is than color chips, and what thing position 9 does not let alone, this road of passing through position 9 with reference to the reference correction of light as light source.When measuring grease color, what thing position 8 does not let alone, and position 9 be the grease sample, this road of passing through position 8 with reference to the reference correction of light as light source.Native system uses bromine tungsten filament lamp, can satisfy the requirement of general measure.
(2) colored thread array CCD
CCD is a photoelectric coupled device, and the sensitive area of colored thread array CCD is to be made of photodiode to resemble quick element array, and three of its picture dots are one group, and each group comprises RGB three looks and resembles quick unit.RGB three looks resemble quick unit and have three spectral response curves, respectively the R in the light beam, G, B composition are carried out sensitization, and the RGB that obtains a picture element represents.
(3) with reference to the light path method
In the present invention, employing is revised the deviation that is caused surely by flashing with reference to the light path method.Can eliminate light source luminescent intensity with reference to the light path method and slowly change the influence that causes.When carrying out machine debugging, behind system stability, gather R, G with reference to light path, B benchmark as light source fluctuation.When the intensity of light source weakens, deduct with reference to the amplitude number percent that fluctuates in the light path with 1 and to remove measured R, G, the B in another road again, be R, the G, the B that eliminate light source fluctuation.When the intensity of light source strengthens, add 1 with the fluctuation amplitude number percent of reference light path and remove measured R, G, the B in another road again, be R, the G, the B that eliminate light source fluctuation.
(4) neural nonlinear fitting
Because it is more inequality than R, G, the B of color chips after opto-electronic conversion to have the oil sample and the Lovibond of identical color and luster, we use neural net method in the present invention, R, G, the B value of oil sample are fitted to the Lovibond that has identical color and luster in database R, G, the B than color chips.Adopt the BP neural network to carry out nonlinear fitting among the present invention.As training sample, after study, this neural network fitting algorithm is applied to other data in the database with R, the G of the existing 20 kinds of oil samples in laboratory, B.After database is proofreaied and correct, behind the R that records a kind of oil sample, G, B, send into neural network earlier and carry out match and obtain R1, G1, B1, in database, retrieve then.During retrieval according to the maximum membership grade principle in the fuzzy theory, in database, seek one group with color R1, G1, the immediate color R2 of B1, G2, B2, i.e. (R1-R2)
2+ (G1-G2)
2+ (B1-B2)
2Value hour, R2, G2, B2 are the Lovibond of this kind oil sample and represent.
Technical scheme of the present invention is: digital photoelectric grease integration color detecting system comprises light source, diffuse reflector, substrate passage, catoptron, electric charge coupling detector C CD, data collecting card, Processing with Neural Network device; Its characteristics are: the light source that adopts symmetric configuration, for system provides even illumination, because the dirt diffused light is produced in the effect of frosted glass, diffuse reflector evenly throws light on, light produces illumination light field uniformly on the substrate passage after the diffuse reflector reflection, the light beam toward mirror of illumination light field enters electric charge coupling detector, data collecting card, Processing with Neural Network device more successively.
The detection method of digital photoelectric grease integration color detecting system, pass through following step:
One. set up Lovibond than color chips database:
Two. set up the oil sample database:
Three. the neural metwork training stage:
Four. test phase:
Beneficial effect of the present invention: the present invention can retrieve the Lovibond of grease faster and represent in database, realizes the robotization of oil sample evaluation, has reduced the cost of tintmeter, has improved accuracy and speed that grease color is measured, has reduced personal error.
Description of drawings
Fig. 1: tintmeter light channel structure synoptic diagram
Fig. 2: the planar structure of substrate;
Fig. 3: Processing with Neural Network device
Among the figure: 1. light source (band frosted glass plate); 2. diffuse reflector; 3. substrate passage; 4. catoptron; 5. electric charge coupling detector C CD; 6. data collecting card; 7. Processing with Neural Network device; 8. reference channel; 9. sample channel.
Embodiment
Below in conjunction with accompanying drawing the specific embodiment of the present invention is described further:
Digital photoelectric grease integration color detecting system comprises light source, diffuse reflector, substrate passage, catoptron, electric charge coupling detector C CD, data collecting card, Processing with Neural Network device; It is characterized in that: the light source 1 that adopts symmetric configuration, for system provides uniform source of light, because the effect of frosted glass produces diffused light, diffuse reflector 2 evenly throws light on, light produces illumination light field uniformly on substrate passage 3 after the diffuse reflector reflection, the light beam toward mirror 4 of illumination light field enters electric charge coupling detector 5, data collecting card 6, Processing with Neural Network device 7 more successively.
Catoptron 4, main effect is to reduce the model machine volume, makes to enter CCD by sample or Lovibond than the light beam of color chips;
Electric charge coupling detector C CD 5 is a linear array colored CCD, and number of pixels is 5340 * 3, and every road dynamic range is 12 bits.
Processing with Neural Network device NN (7) is an integrated circuit, the feed-forward type neural network, and the input node is 3, input quantity is respectively R, the G of sample, the value of B three primary colours, it is three neurons that hiding layer has 7 neurons, output layer, and output is respectively R, the G through match, the value of B three primary colours.
The detection method of digital photoelectric grease integration color detecting system, pass through following step:
One. set up Lovibond than color chips database:
When 1) setting up Lovibond than the color chips database, place Lovibond than color chips at reference channel, sample channel is not placed any substrate;
2) light that sends of light source becomes diffused light behind frosted glass, evenly illumination sample channel and reference channel behind diffuse reflector;
3) light beam through reference channel enters CCD, obtains Lovibond than R, the G of color chips, the value of B three primary colours after the capture card conversion;
4) light signal of collected specimens passage is a benchmark with its light source fluctuation, and Lovibond is summed up in the point that measured value under the same light intensity than the rgb value of color chips;
5) change Lovibond than color chips, repeat above step
Two. set up the oil sample database:
With set up Lovibond seemingly than color chips class database, difference is:
1) do not place substrate in the reference channel, the light signal of this passage is as the qualitative standard of compensation of flashing;
2) oil sample is placed in the sample channel, and CCD gathers the optical information in this passage.
Three. the neural metwork training stage:
1) determines to have the grease of identical color and luster and, set up the training sample database of neural network than color chips;
2),,,, determine network structure up to network convergence according to BP learning method neural network training than the rgb value of color chips desired output as neural network with the rgb value of grease input signal as neural network.
Four. test phase:
1) grease sample to be measured is placed in the sample channel, does not place substrate in the reference channel;
2) utilize CCD to gather optical information in the sample channel, after the capture card conversion, obtain R, the G of this sample, the value of B three primary colours;
3) gathering the light signal of reference channel, is benchmark with its light source fluctuation, and the rgb value of sample is summed up in the point that measured value under the same light intensity;
4) rgb value of the sample after will compensating obtains exporting R1, G1, B1 as the input of neural network after the Neural Network Based Nonlinear mapping;
5) according to the maximum membership grade principle in the fuzzy theory, Lovibond seek in than color chips database one group with color R1, G1, the immediate color R2 of B1, G2, B2, i.e. (R1-R2)
2+ (G1-G2)
2+ (B1-B2)
2Value hour, R2, G2, B2 are the Lovibond of this kind oil sample and represent.
Claims (6)
1. a digital photoelectric grease integration color detecting system comprises light source, diffuse reflector, substrate passage, catoptron, electric charge coupling detector C CD, data collecting card, Processing with Neural Network device; It is characterized in that: the light source (band frosted glass plate) (1) that adopts symmetric configuration, for system provides even illumination, because the effect of frosted glass produces diffused light, diffuse reflector (2) evenly throws light on, light is gone up generation illumination light field uniformly at substrate passage (3) after the diffuse reflector reflection, the light beam toward mirror (4) of illumination light field enters electric charge coupling detector (5), data collecting card (6), Processing with Neural Network device (7) more successively.
2. according to the said digital photoelectric grease integration color detecting system of claim 1, it is characterized in that: substrate passage (3) is divided into reference channel (8) and sample channel (9) two parts, and reference channel (8) is placed Lovibond and compared color chips, sample channel (9) is placed oil sample to be measured; Lovibond is a visual method accepted standard glass sheet group than color chips, totally 12 groups, and comprising three groups of red tab, three groups of erotic film, three groups of blue sheets, its excess-three group is a neutral sheet.
3. according to the said digital photoelectric grease integration color detecting system of claim 1, it is characterized in that: electric charge coupling detector C CD (5) is a linear array colored CCD, and number of pixels is 5340 * 3, and every road dynamic range is 12 bits.
4. according to the said digital photoelectric grease integration color detecting system of claim 1, it is characterized in that: data collecting card (6), be the ADA11GH capture card, its highest sample frequency is 3MHz, and sampling precision is 12Bit, and plate carries the 512Kbyte buffer memory.
5. according to the said digital photoelectric grease integration color detecting system of claim 1, it is characterized in that: Processing with Neural Network device NN (7) is an integrated circuit, the feed-forward type neural network, the input node is 3, input quantity is respectively R, the G of sample, the value of B three primary colours, it is three neurons that hiding layer has 7 neurons, output layer, and output is respectively R, the G through match, the value of B three primary colours.
6. according to the detection method of the said digital photoelectric grease integration color detecting system of claim 1, it is characterized in that it will pass through following step:
One. set up Lovibond than color chips database:
1) place Lovibond than color chips at reference channel, sample channel is not placed any substrate;
2) light that sends of light source becomes diffused light behind frosted glass, evenly illumination sample channel and reference channel behind diffuse reflector;
3) light beam through reference channel enters CCD, obtains Lovibond than R, the G of color chips, the value of B three primary colours after the capture card conversion;
4) light signal of collected specimens passage is a benchmark with its light source fluctuation, and Lovibond is summed up in the point that measured value under the same light intensity than the rgb value of color chips;
5) change Lovibond than color chips, repeat above step.
Two. set up the oil sample database:
With set up Lovibond seemingly than color chips class database, difference is:
1) do not place substrate in the reference channel, the light signal of this passage is as the qualitative standard of compensation of flashing;
2) oil sample is placed in the sample channel, and CCD gathers the optical information in this passage.
Three. the neural metwork training stage:
1) determines to have the grease of identical color and luster and, set up the training sample database of neural network than color chips;
2),,,, determine network structure up to network convergence according to BP learning method neural network training than the rgb value of color chips desired output as neural network with the rgb value of grease input signal as neural network.
Four. test phase:
1) oil sample to be measured is placed in the sample channel, does not place substrate in the reference channel;
2) utilize optical information in the CCD collected specimens passage, after the data collecting card conversion, obtain R, the G of this sample, the value of B three primary colours;
3) gathering the light signal of reference channel, is benchmark with its light source fluctuation, and the rgb value of sample is summed up in the point that measured value under the same light intensity;
4) rgb value of the sample after will compensating obtains exporting R1, G1, B1 as the input of neural network after the Neural Network Based Nonlinear mapping;
5) according to the maximum membership grade principle in the fuzzy theory, Lovibond seek in than color chips database one group with color R1, G1, the immediate color R2 of B1, G2, B2, i.e. (R1-R2)
2+ (G1-G2)
2+ (B1-B2)
2Value hour, R2, G2, B2 are the Lovibond of this kind oil sample and represent.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100133846A CN100458410C (en) | 2006-03-29 | 2006-03-29 | Digital photoelectric grease integration color detecting system and its detecting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100133846A CN100458410C (en) | 2006-03-29 | 2006-03-29 | Digital photoelectric grease integration color detecting system and its detecting method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1821750A true CN1821750A (en) | 2006-08-23 |
| CN100458410C CN100458410C (en) | 2009-02-04 |
Family
ID=36923225
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100133846A Expired - Fee Related CN100458410C (en) | 2006-03-29 | 2006-03-29 | Digital photoelectric grease integration color detecting system and its detecting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100458410C (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102435558A (en) * | 2011-09-13 | 2012-05-02 | 河海大学常州校区 | CO2 sensor based on test solution color measurement and detection method thereof |
| CN102967559A (en) * | 2012-11-26 | 2013-03-13 | 湖南日用化学科学研究所有限公司 | Method for testing color of cloth washing soap or soap base |
| CN103512659A (en) * | 2012-06-19 | 2014-01-15 | 台湾超微光学股份有限公司 | Detachable peripheral device of spectrometer |
| CN107121784A (en) * | 2017-06-14 | 2017-09-01 | 上海仪电物理光学仪器有限公司 | The optical system of automatic lovibond test meter based on area array CCD |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0566015A3 (en) * | 1992-04-14 | 1994-07-06 | Eastman Kodak Co | Neural network optical character recognition system and method for classifying characters in amoving web |
| US6650779B2 (en) * | 1999-03-26 | 2003-11-18 | Georgia Tech Research Corp. | Method and apparatus for analyzing an image to detect and identify patterns |
| CN2498604Y (en) * | 2000-01-25 | 2002-07-03 | 陈慰宗 | Instrument for measuring total content of polarized molecules in edible oil and fat by laser |
-
2006
- 2006-03-29 CN CNB2006100133846A patent/CN100458410C/en not_active Expired - Fee Related
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102435558A (en) * | 2011-09-13 | 2012-05-02 | 河海大学常州校区 | CO2 sensor based on test solution color measurement and detection method thereof |
| CN102435558B (en) * | 2011-09-13 | 2013-11-20 | 河海大学常州校区 | CO2 sensor based on test solution color measurement and detection method thereof |
| CN103512659A (en) * | 2012-06-19 | 2014-01-15 | 台湾超微光学股份有限公司 | Detachable peripheral device of spectrometer |
| CN103512659B (en) * | 2012-06-19 | 2016-05-04 | 台湾超微光学股份有限公司 | Detachable peripheral device of spectrometer |
| CN102967559A (en) * | 2012-11-26 | 2013-03-13 | 湖南日用化学科学研究所有限公司 | Method for testing color of cloth washing soap or soap base |
| CN107121784A (en) * | 2017-06-14 | 2017-09-01 | 上海仪电物理光学仪器有限公司 | The optical system of automatic lovibond test meter based on area array CCD |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100458410C (en) | 2009-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1201016C (en) | Instrument for monitoring polymerase chain reaction of DNA | |
| KR101166180B1 (en) | Methods for controlling one or more parameters of a flow cytometer type measurement system | |
| CN102622746A (en) | Methods and systems for image data processing | |
| CN100339689C (en) | Optical sensor, optical sensing method and storage medium | |
| CN1821750A (en) | Digital photoelectric grease integration color detecting system and its detecting method | |
| CN1366604A (en) | Method and apparatus for detecting mastitis by using visible light and/or near infrared light | |
| CN111157225A (en) | Labview-based EMCCD chip full-performance parameter testing method | |
| CA3143569A1 (en) | Full resolution color imaging of an object | |
| CN1161591C (en) | System for testing optical environment by image processing technique | |
| KR970058431A (en) | Soldering part inspection method and inspection device | |
| FR3075825A1 (en) | METHOD AND SYSTEM FOR IDENTIFYING GRAM TYPE OF A BACTERIUM | |
| CN1131421C (en) | Miniature biochemical analyzer using dual-spectrum detection | |
| FR2905007A1 (en) | QUANTITAVE EVALUATION OF A COLOR FILTER | |
| CN1888833A (en) | Whole-spectrum, double-CMOS linear imaging sensor detector and detecting method thereof | |
| US20200105020A1 (en) | Method of spectral analyzing with a color camera | |
| CA2460266A1 (en) | Method and apparatus for digitizing light measurements by computer control of light source emission | |
| CN1811356A (en) | Spectral measuring method and instrument utilizing planar array photoelectric device | |
| Browne | Imaging and image analysis in the comet assay | |
| CN115144362A (en) | Spectrum model self-adaption method based on standard calibration plate | |
| Vennik et al. | Surface photometry of galaxies in low density regions | |
| CN112326681A (en) | Method for correcting and detecting lens cleanliness by utilizing defocusing difference flat field | |
| CN1289918A (en) | Method and device for multi-channel spectrophotometric detection of rapid microscope | |
| CN109724931B (en) | Real-time calibration spectrum sensing system and spectrum processing method thereof | |
| CN116559119B (en) | Deep learning-based wood dyeing color difference detection method, system and medium | |
| RU2468345C1 (en) | Method for colour classification of objects and electrooptical device for realising said method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090204 Termination date: 20150329 |
|
| EXPY | Termination of patent right or utility model |