CN105956591B - A kind of online hot parts infrared image spectrum sampling Detection method - Google Patents
A kind of online hot parts infrared image spectrum sampling Detection method Download PDFInfo
- Publication number
- CN105956591B CN105956591B CN201610280860.4A CN201610280860A CN105956591B CN 105956591 B CN105956591 B CN 105956591B CN 201610280860 A CN201610280860 A CN 201610280860A CN 105956591 B CN105956591 B CN 105956591B
- Authority
- CN
- China
- Prior art keywords
- parts
- infrared
- image data
- hot parts
- region
- 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.)
- Expired - Fee Related
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 48
- 238000001228 spectrum Methods 0.000 title claims abstract description 32
- 238000005070 sampling Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 26
- 238000000295 emission spectrum Methods 0.000 claims abstract description 22
- 230000003595 spectral effect Effects 0.000 claims abstract description 17
- 238000004458 analytical method Methods 0.000 claims abstract description 15
- 238000004519 manufacturing process Methods 0.000 claims abstract description 14
- 229910052710 silicon Inorganic materials 0.000 claims description 25
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 claims description 24
- 239000004973 liquid crystal related substance Substances 0.000 claims description 24
- 239000010703 silicon Substances 0.000 claims description 24
- 239000000835 fiber Substances 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 7
- 230000008030 elimination Effects 0.000 claims description 2
- 238000003379 elimination reaction Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000002329 infrared spectrum Methods 0.000 abstract description 8
- 238000004566 IR spectroscopy Methods 0.000 abstract description 4
- 238000001816 cooling Methods 0.000 abstract description 3
- 238000010183 spectrum analysis Methods 0.000 abstract description 2
- 238000005259 measurement Methods 0.000 description 4
- 229910001069 Ti alloy Inorganic materials 0.000 description 3
- 239000000956 alloy Substances 0.000 description 3
- 238000003331 infrared imaging Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 239000013307 optical fiber Substances 0.000 description 3
- 229910045601 alloy Inorganic materials 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000701 chemical imaging Methods 0.000 description 1
- 238000010411 cooking Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 229910052738 indium Inorganic materials 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000010813 municipal solid waste Substances 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Classifications
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
-
- 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
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/06—Recognition of objects for industrial automation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Immunology (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Pathology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Spectrometry And Color Measurement (AREA)
Abstract
The present invention provides a kind of infrared image spectrum sampling Detection method of online hot parts, is that infrared image is identified and selected to formula infrared image spectral technique to be fused in the sampling Detection of online hot parts.Using infrared image identification technology, can hot parts multiple on production line be carried out with identification and automated randomized selection;Then selecting type infrared spectrum technology is utilized, detect the emission spectrum of selected hot parts or selected location, and the characteristic peak wavelength of each spectra sample is extracted, operation is modeled to by the long sample sets principal component formed of spectral signature spike, hot parts is carried out in conjunction with spectral analysis technique and is mingled with analysis.This method need not just detect after hot parts are cooling, and multiple infrared images can be carried out with identification and automated randomized detection in the production line, simultaneously part is carried out using infrared spectroscopy to be mingled with analysis, provides a kind of lossless, efficient, non-contacting safety method for the online sampling Detection of hot parts.
Description
Technical field
The present invention relates to infrared image spectrum of use and hot parts detection fields, use choosing more particularly, to a kind of
Select the online hot parts infrared image spectrum sampling Detection method of formula infrared image spectral technique.
Background technique
The lossless non-contact detecting of hot parts and the screening for being mingled with part are the important research sides on alloy production field
To.
With the increase of alloy material usage amount in the industry, importance seems further prominent, and in fusion process
If being mixed into the performance that foreign metal will largely effect on part, or even safety accident can be caused.Such as in aviation flight history, have
Many aircraft accidents are all to lead to power failure since titanium alloy component is broken, and the most fundamental reason of part breaking is just
It is to be mingled with, being mingled with is a kind of volume flaw, and there is when metallic inclusion inside titanium alloy, field trash will cause stress collection
In and in subsequent exercise generate micro-crack, become the tired source under alternating load, thus place start to cause titanium alloy cracking,
It destroys, and eventually leads to material cracks and scrap.Therefore, the detection that is mingled with of alloy part is particularly important.
And when in face of large batch of finished part on production line, sampling Detection is important at this batch of part quality of evaluation
Means, but nowadays most of way is all to have to wait for carrying out composition detection after part is cooling, while may be used also in detection process
It can cause centainly to damage to part, therefore online high temperature non-destructive testing can greatly improve the efficiency of sampling Detection.
And infrared image spectral technique is the new technology of a kind of blending image and spectroscopic data, it can using spectral imaging technology
To identify specific component and detect, or the specific position of part is detected, improves the randomness and detection of sampling
Accuracy.
For online hot parts, part can be visited under the premise of non-contacting using infrared spectrum technology
It surveys, while the effect of spectrum medium wave peak intensity is less, can provide part accurate information is spectral waveform.Therefore these in spectrum
Characteristic makes researcher in the case where that can identify waveform, ignores the influence of luminous intensity or temperature to spectrum, to make
Obtaining infrared spectrum analysis can carry out under a variety of occasions.
Summary of the invention
In order to overcome the above-mentioned deficiencies of the prior art, the present invention proposes a kind of online hot parts infrared image spectrum sampling
Detection method;The safety and nondestructive sampling Detection that online hot parts are realized using the present invention, can be known image using the present invention
Do not blend with infrared spectrum detection, identify whole hot parts on production line on the image, then at random to some zero
Part or specific position carry out infrared spectrum detection and carry out folder analysis.It can effectively solve traditional sampling Detection method to be difficult in height
It is detected in warm situation, specific position can not be detected and the problems such as damage is caused to part.
To achieve the goals above, technical scheme is as follows:
A kind of infrared image spectrum sampling Detection method of online hot parts, comprising:
S201. production line Parts Recognition region is planned, and dividing identification region equally is several search windows;
S202. before part does not enter identification region, bias light image data is acquired;
S203. first part enters identification region R, acquires the image data of first part;
S204. more a parts enter identification region, obtain the image information of all parts in identification region;Integral calculation is each
The grey level histogram of search window, and the image information of first part of integral calculation and the image data of bias light are as mark
Standard matches the highest region of image data similarity after filtering off bias light with first part in each search window and makees
For parts profile, and whole parts profiles are successively numbered;
S205. a dash number is chosen at random, generates the mask artwork of parts profile ROI, by recording the part feature
Emission spectrum;
S206. judge whether the part exceeds the search window or whether the sample number of part exceeds the acquisition upper limit;If all
It is no, then carry out next step S207;If one of them is, step S208 is carried out;
S207. it is matched again in the search window according to the image data of the part, finds out the profile of the part
And continue to carry out spectrographic detection to this part, and form the characteristic emission spectrum collection of the part, enter back into S206;
S208. it removes the number of the part and abandons the collecting work to the part;
S209. it is long as principal component sample sets to extract spectral signature spike;
S210. characteristic peak wavelength value and number of wavelengths are subjected to modeling analysis, judge whether part is mingled with.
Preferably, before part does not enter cog region in step S202, bias light figure is acquired using infrared area array CCD camera
As data;First part enters identification region R in step S203, acquires first part using infrared area array CCD camera
Image data;Multiple parts enter identification region in step S204, are obtained in identification region and are owned using infrared area array CCD camera
The image information of part.
Preferably, in step S205, processor randomly chooses a part, and controls the reflective infrared liquid crystal on silicon of LCOS
The mask artwork of parts profile ROI is generated, then the part feature emission spectrum is recorded by fiber spectrometer;
It in step S207, is matched again in the search window according to the image data of the part, finds out the part
Profile and be input in the reflective infrared liquid crystal on silicon of LCOS, fiber spectrometer continue to this part carry out spectrographic detection, and
Form the characteristic emission spectrum collection of the part.
Preferably, the pixel length-width ratio and infrared area array CCD phase of the reflective infrared liquid crystal on silicon working region the LCOS
The pixel length-width ratio of machine working region is identical, and the reflective infrared liquid crystal on silicon working region number of pixels of LCOS is greater than infrared surface
The number of pixels of array CCD camera working region.
Preferably, in step 203, the reflective infrared liquid crystal on silicon of LCOS is in shading status, at this point, light cannot pass through
The reflective infrared liquid crystal on silicon of LCOS is received by fiber spectrometer.
Preferably, it is specifically to take single order to spectrum that spectral signature spike length is extracted in step 209 as principal component sample sets
The null abscissa wavelength X of derivative valueiAs principal component sample, i=1,2 ..., N work asWhen, wavelength X1、
λ2......λNAs principal component sample sets;
Characteristic peak wavelength value and number of wavelengths are carried out modeling analysis by step S210, judge the judgement item whether part is mingled with
Part are as follows:
1.
2. working asWhen,
Wherein N is the characteristic wavelength number of part to be measured,For the average characteristics number of wavelengths of part each in sample sets, λi
For the ith feature wavelength of part to be measured,For sample sets ith feature wavelength average value, K is variance coefficient of determination, and M is
The total number of all parts in sample sets,λ ji, it is the ith feature wavelength of j-th of sample in sample sets.
As a result, on the basis of image recognition, the present invention, which is detected and is mingled with the emission spectrum to hot parts, to be determined as
Technological core distinguishes whether part is mingled with this from the otherness of part emission spectrum.And image spectrum detects and divides
Analysis technology can realize the measurement of automation with automatic identification spectrum and image data, provide technical support for online detection,
It is a kind of increasingly important research means in industrial production.
Efficiently online hot parts sampling Detection method provided by the invention, identifies and selects the infrared figure of formula for infrared image
It is fused to as spectral technique in the sampling Detection of online hot parts.It, can be to more on production line using infrared image identification technology
A hot parts carry out identification and automated randomized selection;Then selecting type infrared spectrum technology is utilized, selected hot parts are detected
Or the emission spectrum of selected location, and the characteristic peak wavelength of each spectra sample is extracted, to what is be made of spectral signature spike length
Sample sets principal component models operation, carries out hot parts in conjunction with spectral analysis technique and is mingled with analysis.This method need not wait until high temperature
It is just detected after part is cooling, and multiple infrared images can be carried out with identification and automated randomized detection in the production line, simultaneously
Part is carried out using infrared spectroscopy to be mingled with analysis, for hot parts online sampling Detection provide it is a kind of lossless, efficient, non-
The safety method of contact.
Compared with prior art, the beneficial effect of technical solution of the present invention is: the present invention utilizes infrared image spectral measurement
The sampling Detection to online hot parts is realized with analytical technology, online multiple hot parts can be identified, and with
Machine chooses some part or some position carries out infrared spectrometry, while the characteristic wavelength for extracting spectrum can press from both sides part
Miscellaneous analysis.Compared with traditional sampling Detection technology, this method, which is realized, detects the non-contact safety on line of hot parts, detection
Process realizes automatic identification and chooses to improve detection speed, can also choose fixed part position detection to improve
Accuracy rate, while part will not be caused to damage using infrared spectrum analysis.
Detailed description of the invention
Fig. 1 is a kind of infrared image spectrum sampling Detection method flow diagram of online hot parts provided by the invention.
Fig. 2 is a kind of infrared image spectrum sampling Detection method system figure of online hot parts provided by the invention.
Fig. 3 is the physical layout drawings of hot parts identification region.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent;In order to better illustrate this embodiment, attached
Scheme certain components to have omission, zoom in or out, does not represent the size of actual product;
To those skilled in the art, it is to be understood that certain known features and its explanation, which may be omitted, in attached drawing
's.The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
The present invention realizes the purpose of the infrared image spectrum sampling Detection of online hot parts, can be to online multiple high temperature
Part carries out image recognition and randomly selects part progress infrared spectrum detection and analysis.
It is provided below according to Fig. 1, to illustrate the infrared image spectrum sampling Detection side of online hot parts of the invention
Method.
S201: in production line simulation Parts Recognition region, and several search windows are bisected into;
S202: before part does not enter cog region, bias light image data is acquired;
S203: when first part enters identification region R, the image data of first part is acquired;
S204: multiple parts enter identification region, obtain the image information of all parts in identification region;Integral calculation is each
The grey level histogram of search window, and the image information of first part of integral calculation and the image data of bias light are as mark
Standard matches the highest region of image data similarity after filtering off bias light with first part in each search window and makees
For parts profile, and whole parts profiles are successively numbered;
S205: computer random chooses a dash number, and computer, which controls LCOS and generates, only has parts profile ROI's
Mask artwork records the part feature emission spectrum by fiber spectrometer;
S206: judge whether the part exceeds search window or whether the sample number of part exceeds the acquisition upper limit;If all no,
Then carry out next step step S207;If one of them is, step S208 is carried out;
S207: being matched in the search window according to the image data of the part again, finds out the profile of the part
And be input in LCOS, fiber spectrometer continues to carry out spectrographic detection to this part, and forms the characteristic emission spectrum of the part
Collection, enters back into step S206;
S208: removing the number of the part and abandons the collecting work to the part;
S209: it is long as principal component sample sets to extract spectral signature spike;
S210: characteristic peak wavelength value and number of wavelengths are subjected to modeling analysis, judge whether to be mingled with.
Above-mentioned online hot parts infrared image spectrum sampling Detection method is based on selecting type infrared image spectral technique
Online hot parts sampling Detection method, specifically includes the following steps:
(1) the region R of Parts Recognition is cooked up in the production line, and R is divided into H search window Si, i=1,2,
3 ..., H set the characteristic spectrum sample acquisition upper limit of each part as M;
(2) the information light in identification region R enters infrared beam-dividing cube through infrared imaging lens group and is split, and transmits
Light beam focuses on infrared area array CCD camera through the second convergent lens, obtains the image information of bias light in identification region R, reflection
Light beam is assembled through the first polarizing film (polarizer slice), the reflective infrared liquid crystal on silicon of LCOS, the second polarizing film (checking bias slice) and first
Lens enter in fiber spectrometer after converging to fiber coupler, obtain the characteristic spectrum information in the R of Parts Recognition region;It is infrared
Area array CCD camera and the reflective infrared liquid crystal on silicon of LCOS are connected with computer;
(3) at the T moment, first hot parts enters identification region R, and infrared area array CCD camera obtains high temperature in region
The image information of part;
(4) at the T+m* △ T moment, multiple hot parts enter identification region R;
(5) infrared area array CCD camera obtains the image information of hot parts in identification region, and computer is by each search window
SiInterior image data carries out integral operation and obtains grey level histogram T respectivelyi, i=1,2,3 ..., H, and integral calculation first
Part image data T1With the image data T of bias lightRIt as standard, then asks poor, enables Δ Ti=Ti-TRAs elimination bias light
The image data of i-th of part, Δ T1=T1-TRAs the image data for first part for filtering off bias light, it is according to Pasteur
Counting method is in each search window SiImage data Δ TiIn select and Δ T1The highest region of similarity as parts profile,
And L is successively numbered to whole parts profilesi, i=1,2,3 ..., H;
(6) computer random chooses a dash number W, and computer controls the reflective infrared liquid crystal on silicon of LCOS and generates only
There is the mask artwork of this target part profile ROI, the part feature emission spectrum F at the moment is successively recorded by fiber spectrometerLw,
(7) judge whether each part exceeds the field of search, and judge whether the sample number of each part exceeds the acquisition upper limit simultaneously.
If without departing from the field of search and not the sample number of part is not more than the acquisition upper limit to part, in the search window according to this zero
Image data Δ T of the part at the T+m* △ T momentwImages match is carried out with Pasteur's Y-factor method Y, finds out the profile of the part again simultaneously
It is input in the reflective infrared liquid crystal on silicon of LCOS, fiber spectrometer continues to carry out this part spectrographic detection, and by multiple same
The data of one part form the characteristic emission spectrum collection of each part, duplicate step of laying equal stress on;If part occur crosses the field of search, or works as
When single part emission spectrum sample number occur and being more than the acquisition upper limit, computer system removes the number of the part and abandons pair
The collecting work of the part, and the null characteristic wavelength λ of first derivative values is taken to its characteristic emission spectrumiAs principal component, i
=1,2,3 ..., H establish specific judgement and are mingled with model and show judgement result.
A preferable embodiment of the invention is provided below according to Fig. 2, it is real to illustrate system structure characteristic of the invention
The infrared image spectrum sampling Detection method of present line hot parts, rather than range for the purpose of limiting the invention.
1- infrared imaging lens group, the infrared beam-dividing cube of 2-, 3, the first and second polarizing film of 5-, 4-LCOS it is reflective infrared
Liquid crystal on silicon, 6, the first and second convergent lens of 10-, 7- fiber coupler, 8- optical fiber, 9- fiber spectrometer, the infrared area array CCD of 11-
Camera, 12- computer.
As shown in Fig. 2, in the present embodiment, a kind of infrared image spectrum sampling Detection system of online hot parts includes red
Outer imaging lens group 1, infrared beam-dividing cube 2, the first polarizing film 3, the second polarizing film 5, the reflective infrared liquid crystal on silicon of LCOS
4, the first convergent lens 6, the second convergent lens 10, fiber coupler 7, optical fiber 8, fiber spectrometer 9, infrared area array CCD camera
11, computer 12.
In the present embodiment, infrared 2 sensitive wavelength of beam-dividing cube is 780nm-1550nm, and reflection and transmission ratio is 1;Optical axis is inclined
It moves less than 10 points.
In the present embodiment, infrared 11 sensitive wavelength of area array CCD camera is 780nm-1550nm, and resolution ratio is 680 × 520,
2/3 inch, Pixel Dimensions are that 6.45um × 6.45um image frame per second is set as 15fps, i.e., every frame duration T=67ms.
In the present embodiment, the spectral measurement ranges of fiber spectrometer 9 are 780nm-1650nm, optical resolution 10nm,
The time of integration 1ms to more than 65s, and the time of integration is initialized as 67ms.
In the present embodiment, the reflection wavelength of the reflective infrared liquid crystal on silicon 4 of LCOS is 780nm-1550nm, and resolution ratio is
800×600.Since the long and wide number of pixels of the reflective infrared liquid crystal on silicon 4 of LCOS is all larger than infrared area array CCD camera 11,
Thus by manually calibrating, make the corresponding one infrared area array CCD camera 11 of the pixel of the reflective infrared liquid crystal on silicon 4 of each LCOS
Pixel, to realize that any pixel carries out spectral measurement in system energy correspondence image.
In the present embodiment, image light is divided into two bundles after the collimation of infrared imaging lens group 1 through infrared beam-dividing cube 2, and one
Infrared area array CCD camera 11 is imaged in through the second convergent lens 10 after beam transmission, computer 12 identifies entirely according to image data
Portion's parts profile is then random to be selected one and is input in the reflective infrared liquid crystal on silicon 4 of LCOS.It is passed through after another light beam reflection
First polarizing film 3 (polarizer) enters the reflective infrared liquid crystal on silicon 4 of LCOS, and light is by modulation output to 5 (analyzing of the second polarizing film
Device) after be transmitted on fiber coupler 7 by the first convergent lens 6, optical information is transferred to fiber spectrometer 9 along optical fiber 8, output
Spectral information is analyzed by computer 12 again.This system has structure simple, operates more safely and flexibly feature.
The infrared spectroscopy sampling Detection method flow of online hot parts is as follows:
Step 1: setting system is 30 to the acquisition upper limit of the sample number of hot parts emission spectrum.
Step 2: cooking up Parts Recognition region R on production line, determines the starting line of Parts Recognition region R to finishing line
Distance be 2.5 meters, and identification region is bisected into 5 search windows, the setting of identification region and search window such as Fig. 3 institute
Show.
If production line part movement speed is 0.1m/s, from the search window is entered to the region is left, can obtain complete
The frame number of parts profile is about 75 frames, that is, ignores under acquisition upper limit case, can obtain spectra sample number and be up to 75.
Step 3: before hot parts do not enter production line, infrared area array CCD camera 11 obtains the image data of environment light;
Step 4: first hot parts enters identification region R, and infrared area array CCD camera 11 is obtained by light signal collection
Obtain the image information of hot parts in region;
Step 5: multiple hot parts enter identification region R, and computer 12 is known by infrared area array CCD camera 11
The image information of hot parts in other region, computer 12 is according to 5 search window SiInterior image data, each window of integral calculation
The grey level histogram T of mouthiAnd the image data T of the 1st part1With the image data T of bias lightR, then ask poor, enable Δ Ti=
Ti-TRAs the image data for i-th of part for filtering off bias light, Δ T1=T1-TRAs first part for filtering off bias light
Image data, according to Pasteur's coefficient method in each search window SiImage data Δ TiIn select and Δ T1Similarity is most
L is successively numbered as parts profile, and to whole parts profiles in high regioni, i=1,2,3 ..., 5;
Step 6: computer 12 chooses a dash number 3 at random, and computer 12 controls the reflective infrasil base fluid of LCOS
Crystalline substance 4 sequentially generates the mask artwork of only No. 3 parts profile ROI, is sent out by the part feature that fiber spectrometer 9 successively records the moment
Penetrate spectrum FL3。
Step 7: judging whether No. 3 parts exceed the 3rd field of search, and judges whether the sample number of part exceeds simultaneously
Acquire the upper limit.If without departing from the 3rd field of search and not the sample number of part is not more than the acquisition upper limit to No. 3 parts, the 3rd
The resulting image data Δ T for filtering off bias light is calculated in step 5 according to the part in a search window3With Pasteur's coefficient
Method carries out images match, finds out the profile of the part again and is input in the reflective infrared liquid crystal on silicon 4 of LCOS, fiber spectrum
Instrument 9 continues to carry out spectrographic detection to this part, forms the characteristic emission spectrum collection of the part, duplicate step of laying equal stress on;If No. 3 parts
Cross the 3rd field of search or when No. 3 part emission spectrum sample numbers are more than the acquisition upper limit 30,12 system of computer remove this zero
The collecting work to the part is numbered and abandoned to part, enters step the acquisition of the five next parts of carry out.
Step 8: when all parts leave data collection area, or when all part emission spectrum sample sets numbers are more than to adopt
When collecting the upper limit, then stop carrying out data acquisition to this batch of part completely.
Modeling for part is classified as calibration set and forecast set after obtaining characteristic emission spectrum.Single order is taken to spectrum
The null abscissa wavelength X of derivative valuei(i=1,2,3...) is used as principal component, that is, works asWhen, λ1、λ2......λN's
Value is characterized wavelength as principal component, and the characteristic wavelength of calibration set is taken to establish the sample sets of part principal component modeling.
Forecast set meets one of the following conditions and is judged to being mingled with:
1.
2. working asWhen,
Wherein N is the characteristic wavelength number of part to be measured,For the average characteristics number of wavelengths of part each in sample sets, λi
For the ith feature wavelength of part to be measured,For sample sets ith feature wavelength average value, K is variance coefficient of determination, and M is
The total number of all parts in sample sets,λ jiFor the ith feature wavelength of j-th of sample in sample sets.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (6)
1. a kind of infrared image spectrum sampling Detection method of online hot parts characterized by comprising
S201. production line Parts Recognition region is planned, and dividing identification region equally is several search windows;
S202. before part does not enter identification region, bias light image data is acquired;
S203. first part enters identification region R, acquires the image data of first part;
S204. more a parts enter identification region, obtain the image information of all parts in identification region;Integral calculation is respectively searched for
The grey level histogram of window, and the image information of first part of integral calculation and the image data of bias light be as standard, then
Ask poor, show that all parts filter off the image data of bias lights and first part filters off the image data of bias light, according to bar
It is matched in the image of elimination bias light of family name's coefficient method in each search window after filtering off bias light with first part
The highest region of image data similarity be successively numbered as parts profile, and to whole parts profiles;
S205. a dash number is chosen at random, generates the mask artwork of parts profile ROI, by recording part feature transmitting
Spectrum;
S206. judge whether the part exceeds the search window or whether the sample number of part exceeds the acquisition upper limit;If all no,
Carry out next step S207;If one of them is, step S208 is carried out;
S207. matched again in the search window according to the image data of the part, find out the profile of the part and after
It is continuous that spectrographic detection is carried out to this part, and the characteristic emission spectrum collection of the part is formed, enter back into S206;
S208. it removes the number of the part and abandons the collecting work to the part;
S209. it is long as principal component sample sets to extract spectral signature spike;
S210. characteristic peak wavelength value and number of wavelengths are subjected to modeling analysis, judge whether part is mingled with.
2. method according to claim 1, which is characterized in that before part does not enter cog region in step S202, use is red
Outer area array CCD camera acquires bias light image data;First part enters identification region R in step S203, using infrared surface
Array CCD camera acquires the image data of first part;Multiple parts enter identification region in step S204, using infrared surface battle array
CCD camera obtains the image information of all parts in identification region.
3. method according to claim 1 or claim 2, which is characterized in that in step S205, processor randomly chooses a part,
And control the mask artwork that the reflective infrared liquid crystal on silicon of LCOS generates parts profile ROI, then by fiber spectrometer record this zero
Part characteristic emission spectrum;
It in step S207, is matched again in the search window according to the image data of the part, finds out the wheel of the part
Exterior feature is simultaneously input in the reflective infrared liquid crystal on silicon of LCOS, and fiber spectrometer continues to carry out spectrographic detection to this part, and is formed
The characteristic emission spectrum collection of the part.
4. method according to claim 3, which is characterized in that the picture of the reflective infrared liquid crystal on silicon working region LCOS
Plain length-width ratio is identical as the pixel length-width ratio of infrared area array CCD camera working region, and the reflective infrared liquid crystal on silicon work of LCOS
Make the number of pixels that area pixel number is greater than infrared area array CCD camera working region.
5. method according to claim 1, which is characterized in that in step 203, the reflective infrared liquid crystal on silicon of LCOS, which is in, to be hidden
Light state, at this point, light cannot be received by the reflective infrared liquid crystal on silicon of LCOS by fiber spectrometer.
6. method according to claim 1, which is characterized in that it is long as principal component to extract spectral signature spike in step 209
Sample sets are specifically to take the null abscissa wavelength X of first derivative values to spectrumiAs principal component sample, i=1,2 ...,
N works asWhen, wavelength X1、λ2......λNAs principal component sample sets;
Characteristic peak wavelength value and number of wavelengths are carried out modeling analysis by step S210, judge the decision condition whether part is mingled with
Are as follows:
2. working asWhen,
Wherein N is the characteristic wavelength number of part to be measured,For the average characteristics number of wavelengths of part each in sample sets, λiFor to
The ith feature wavelength of part is surveyed,For sample sets ith feature wavelength average value, K is variance coefficient of determination, and M is sample sets
In all parts total number,λ jiFor the ith feature wavelength of j-th of sample in sample sets.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610280860.4A CN105956591B (en) | 2016-04-28 | 2016-04-28 | A kind of online hot parts infrared image spectrum sampling Detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610280860.4A CN105956591B (en) | 2016-04-28 | 2016-04-28 | A kind of online hot parts infrared image spectrum sampling Detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105956591A CN105956591A (en) | 2016-09-21 |
CN105956591B true CN105956591B (en) | 2019-03-29 |
Family
ID=56916440
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610280860.4A Expired - Fee Related CN105956591B (en) | 2016-04-28 | 2016-04-28 | A kind of online hot parts infrared image spectrum sampling Detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105956591B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106706133B (en) * | 2016-12-31 | 2017-09-29 | 华中科技大学 | One kind point plaque-like object's pose estimation method and system |
CN110501914B (en) * | 2018-05-18 | 2023-08-11 | 佛山市顺德区美的电热电器制造有限公司 | Security monitoring method, equipment and computer readable storage medium |
CN112712554B (en) * | 2020-12-08 | 2022-08-26 | 上海交通大学烟台信息技术研究院 | Method for extracting central line of laser stripe on surface of semitransparent Lambert surface |
CN113640445A (en) * | 2021-08-11 | 2021-11-12 | 贵州中烟工业有限责任公司 | Characteristic peak identification method based on image processing, computing equipment and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103534581A (en) * | 2011-03-21 | 2014-01-22 | 费德罗-莫格尔公司 | Multi-spectral imaging system and method of surface inspection therewith |
CN103925999A (en) * | 2014-05-06 | 2014-07-16 | 中山大学 | Image spectrum detection method and system |
CN104897637A (en) * | 2015-03-19 | 2015-09-09 | 复旦大学 | Food safety detection apparatus and detection method based on Raman scattering |
CN105043544A (en) * | 2015-05-05 | 2015-11-11 | 中山大学 | Image spectrum tracking method and system |
CN105181595A (en) * | 2015-05-25 | 2015-12-23 | 中山大学 | Plant leaves health monitoring system based on visual image spectrum detection technology |
-
2016
- 2016-04-28 CN CN201610280860.4A patent/CN105956591B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103534581A (en) * | 2011-03-21 | 2014-01-22 | 费德罗-莫格尔公司 | Multi-spectral imaging system and method of surface inspection therewith |
CN103925999A (en) * | 2014-05-06 | 2014-07-16 | 中山大学 | Image spectrum detection method and system |
CN104897637A (en) * | 2015-03-19 | 2015-09-09 | 复旦大学 | Food safety detection apparatus and detection method based on Raman scattering |
CN105043544A (en) * | 2015-05-05 | 2015-11-11 | 中山大学 | Image spectrum tracking method and system |
CN105181595A (en) * | 2015-05-25 | 2015-12-23 | 中山大学 | Plant leaves health monitoring system based on visual image spectrum detection technology |
Non-Patent Citations (2)
Title |
---|
一种面向PX装置成品杂质含量检测的拉曼光谱分析法;卢艺楠 等;《光散射学报》;20131231;第25卷(第4期);全文 |
基于可见短波近红外高光谱图像的梳棉杂质关键波长的优选;郭俊先 等;《新疆农业科学》;20160229;第53卷(第2期);全文 |
Also Published As
Publication number | Publication date |
---|---|
CN105956591A (en) | 2016-09-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105956591B (en) | A kind of online hot parts infrared image spectrum sampling Detection method | |
CN104914111B (en) | A kind of steel strip surface defect online intelligent recognition detecting system and its detection method | |
CN103439342B (en) | Based on the Infrared Non-destructive Testing method of thermal map temporal aspect | |
US5654977A (en) | Method and apparatus for real time defect inspection of metal at elevated temperature | |
Schmugge et al. | Crack segmentation by leveraging multiple frames of varying illumination | |
WO2019039313A1 (en) | Real-time spectral analysis through high-speed spectral classification | |
CN112461892B (en) | Infrared thermal image analysis method for nondestructive detection of composite material defects | |
CN104897583B (en) | A kind of micro bloodstain the cannot-harm-detection device and method | |
CN105974120B (en) | Automatic detection device and method for C-reactive protein chromaticity | |
CN106442525A (en) | Online detecting method for wizened defects of insides of walnuts | |
CN106446799A (en) | Thermal imaging target identification method and apparatus | |
CN107679569A (en) | Raman spectrum substance automatic identifying method based on adaptive hypergraph algorithm | |
CN102621116A (en) | Detection device and method for quality evaluation of American ginseng | |
CN104656100A (en) | Line-scanning hyperspectral real-time anomaly detection method and system | |
CN106022365B (en) | Surface defect depth estimation method based on data fusion and RBF neural | |
CN108180871B (en) | method for quantitatively evaluating composite insulator surface pulverization roughness | |
CN109001182A (en) | The Raman spectrum non-destructive determination method of alcohol content in closed container | |
CN106323937B (en) | A kind of the Pubei oilfield spectrum structure and discrimination method of high sense | |
CN112818806A (en) | Transformer substation inspection robot auxiliary navigation method based on deep learning | |
JP2000258584A (en) | Site inspection system | |
CN115356479B (en) | Gold immunochromatography detection method and system | |
JPH10185911A (en) | Device for analyzing cell and method therefor | |
CN105606544A (en) | Nondestructive detection method of insect bodies of Cordyceps sinensis | |
CN102507462B (en) | Hyperspectral-analysis-based copper quality detection method and system | |
US20140125811A1 (en) | System, method and computer program product for identifying chemical composition of optically lifted latent prints and contaminants |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190329 Termination date: 20210428 |
|
CF01 | Termination of patent right due to non-payment of annual fee |