CN100484191C - Method for generating image without time resolution limit based on light machine scanning signal - Google Patents
Method for generating image without time resolution limit based on light machine scanning signal Download PDFInfo
- Publication number
- CN100484191C CN100484191C CNB2004100840424A CN200410084042A CN100484191C CN 100484191 C CN100484191 C CN 100484191C CN B2004100840424 A CNB2004100840424 A CN B2004100840424A CN 200410084042 A CN200410084042 A CN 200410084042A CN 100484191 C CN100484191 C CN 100484191C
- Authority
- CN
- China
- Prior art keywords
- time
- pixel
- curve
- image
- value
- 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
Images
Landscapes
- Image Processing (AREA)
- Radiation Pyrometers (AREA)
Abstract
This invention relates to a method for generating picture signals acquired by an optical computer scan imaging system to a without limitation of time resolution and no time difference among picture elements characterizing in utilizing the correlation of each point time and number in the optical computer scan image frames to process the noise reduction and fit the time to find an optimum parameter vector and divides all picture elements into different regions based on the character by judging distance between each picture element value and fit curves then to iterate and fit them many times so as to find out different regions corresponding fit parameter vectors.
Description
Technical field
The present invention relates to the method that a kind of picture signal that the optical mechaical scanning imaging system is gathered generates image without time resolution limit, is a kind of improvement to traditional optical mechaical scanning imaging system.
Background technology
The optical mechaical scanning imaging system is a kind of more outmoded data acquisition and imaging system, be characterized in reading one by one the picture signal that information on the target obtains target, and be stored in the memory, the display unit that then picture signal in the memory is sent to system shows.According to the shape of optical mechaical scanning imaging system scanning field of view and the arrangement mode of detector, multiple track while scan can be arranged: line scanning, television scanning and petal scanning etc., but no matter take which kind of scan mode, the data point of its collection all is discrete, this has just caused in the picture signal of optical mechaical scanning imaging system collection, and different pixel acquisition time points are inconsistent.So, the picture frame of optical mechaical scanning imaging system collection record be the change information of target in a period of time.
Because the defective on the optical mechaical scanning imaging system principle, there are two bigger problems in the view data that it is gathered: 1, frequency acquisition is limited; 2, the acquisition time of different pixels is inconsistent on the same frame.Particularly second problem brought bigger difficulty for the data of handling the collection of optical mechaical scanning imaging system, wherein the situation of mixing during especially for the information of same in different characteristic zone.But the acquisition time of adjacent picture elements is quite short at interval in the optical mechaical scanning signal, even under the very low situation of the scanning frequency of optical mechaical scanning imaging system, that is to say that the optical mechaical scanning imaging system in fact provides the signal fragment of some high time resolutions.This provides the foundation for the image that the picture signal by the collection of optical mechaical scanning imaging system generates no temporal resolution restriction.
Summary of the invention
The objective of the invention is to overcome the weakness of existing optical mechaical scanning imaging system, bring into play its strong point, finally generate the picture signal of no temporal resolution restriction by the optical mechaical scanning signal.The present invention realizes purpose of the present invention by the optical mechaical scanning signal is handled.Major measure has the picture signal of pair optical mechaical scanning to carry out the noise reduction sonication, pixel in the image is carried out correlation analysis, determine the Changing Pattern that pixel has, distinguish the residing characteristic area of pixel, quantize the Changing Pattern of pixel with fitting method, thereby generate the digital picture of no temporal resolution restriction, wherein, noise reduction sonication and correlation analysis are the keys of whole process, by this two steps processing, just can tell the residing zones of different of target pixel, thereby reach the purpose that improves the image temporal resolution.
To achieve these goals, a kind of the optical mechaical scanning signal is handled the method that can generate image without time resolution limit thereby the invention provides, concrete grammar is as follows:
A kind of method based on optical mechaical scanning signal generation image without time resolution limit comprises:
The optical mechaical scanning imaging system reads the information that detects on the target one by one and obtains its picture signal, and picture signal is stored in the known memory frame by frame, and wherein, the information of each pixel point comprises temperature value and corresponding acquisition time thereof;
Display unit is converted to image with the picture signal in the known memory;
What the present invention did was improved to before the picture signal in the known memory being delivered to the display unit demonstration, pass through following treatment step successively:
(1) reads in the temperature value and the time corresponding value thereof of all pixels on a certain frame;
(2) picture signal is carried out the noise reduction sonication;
(3) all pixels in the image are carried out correlation analysis, obtain the variation rule curve of all pixels on a certain frame, the i.e. temperature-time curve of match by general approximating method;
(4) calculate each pixel value to matched curve apart from d, according to each pixel value to apart from d each pixel point being carried out characteristic area and divides of matched curve, according to the maximum allowable range d of setting
mThe different situations of value and d magnitude relationship are divided into three classes with all pixel points:
The first kind: | d|＜d
m, these points belong to same characteristic area;
Second class: d〉d
m, forward step (3) to, the pixel point in the class is carried out process of fitting treatment once more, obtain a new temperature-time curve, then to the processing of pixel point repeating step (4) in the class;
The 3rd class: d＜-d
m, forward step (3) to, the pixel point in the class is carried out process of fitting treatment once more, obtain a new temperature-time curve, then to the processing of pixel point repeating step (4) in the class;
Till the classification results that obtains all was first kind pixel point, so far, the characteristic area partition process finished;
(5) all pixels in each characteristic area are carried out process of fitting treatment, the pixel that is in the same area has identical Changing Pattern, as long as promptly obtain a fitting parameter vector, and just can this regional variation characteristic of unique description;
(6) according to the fitting parameter vector of each characteristic area, substitution time t promptly can calculate the temperature value of any all pixels of the moment in the pairing time period of this two field picture;
At last the data image signal that generates is delivered to the display unit demonstration and obtained digital picture.
The noise reduction sound process of step in the such scheme (2) is utilized the scheme of general wicket weighting smothing filtering, and the design of coefficients should make the horizontal direction weight factor big, and vertically weight factor is little.
Dependency analysis process in the step (3) requires the way of elder generation by match, obtain the variation rule curve of all pixels on a certain frame, calculate the distance of each pixel value then to curve, divide characteristic area under each pixel according to the size of distance, all kinds of pixels carry out the match second time after separating again, divide again, repeat above process, up to the maximum allowable range d of distance less than the criterion setting
mDefinition pixel value is d to the distance of curve f, and in plane right-angle coordinate, on curve, d is a positive number, otherwise is negative as if the pixel value, and the maximum allowable range of establishing the criterion requirement is d
m
In the area dividing process of step (4), the desired maximum allowable range d of criterion
mThe set point size should be less than the greatest measure difference of zones of different greater than the maximum noise of system.
In step (5) quantizes same characteristic area in the process of pixel Changing Pattern, the pixel that is in the same area is adopted the way of match, obtain a fitting parameter vector, by this fitting parameter is vectorial just can this regional variation characteristic of unique description.
In the process of the infinitely-great digital picture of step (6) rise time resolution, owing to obtained the fitting parameter vector of each characteristic area, as long as provide time value, just can obtain the temperature value of all pixels of any time, thereby generate the digital picture of no temporal resolution restriction.
Compare with the focal-plane array (FPA) scanning imaging system, the optical mechaical scanning imaging system is the relatively backward low side scanning system of a kind of scan mode, but Chinese domestic use optical mechaical scanning imaging system is carried out the also quite a lot of data acquisition.Giving full play to the potentiality of existing equipment, the signal of optical mechaical scanning imaging system collection is generated as the image that no temporal resolution limits, is exactly the target that the present invention lays siege to.
The present invention utilizes the time of each pixel in the optical mechaical scanning picture frame and the correlation of pixel value, by carrying out the noise reduction sonication earlier, reduce the influence of systematic error to result, carry out curve fitting then, find out optimum coefficient, by judging the distance between each pixel value and the matched curve, all pixels are carried out piecemeal, divide different characteristic areas with this, and then these pixels are carried out match, thereby find out one group of fitting parameter vector of zones of different correspondence, the substitution time can obtain the temperature value of all pixels of any time.Wherein, pass through fitting method, pixel is carried out piecemeal, be at the shortcoming of optical mechaical scanning imaging system and the effective measures of taking, it has made full use of the advantage of optical mechaical scanning imaging system institute images acquired, has reduced the influence of optical mechaical scanning imaging system self shortcoming to its image data simultaneously again greatly.
D among the present invention
mBe that the maximum allowable range that criterion is set is the parameter of a key, it has great influence to the number of regions and the effect of dividing.When calculating, can utilize a circulation checking to determine best d by program
mValue.As the d that determines a best
mAfter, just can successfully finish the area dividing of pixel, calculate each regional fitting parameter vector then, just can obtain the numerical value of any time pixel after the substitution time.
Outstanding advantage of the present invention:
(1) overcome the optical mechaical scanning imaging system to the inconsistent problem of the acquisition time of each pixel.The present invention has obtained the fitting parameter vector of all pixels at last by methods such as matches, like this, as long as provide time parameter, just can obtain the numerical imaging of all pixel points of any time.
(2) the lower optical mechaical scanning image of scanning frequency is generated the image that no temporal resolution limits, improved the time and the spatial resolution of image greatly.
Description of drawings
Fig. 1: the flow chart that the picture signal that optical mechaical scanning is obtained is handled;
Fig. 2: the view data that the optical mechaical scanning imaging system is gathered, the unprocessed display unit of directly delivering to shows observed result;
Fig. 3: pixel point actual value distribution situation and the temperature-time graph that carries out obtaining after the match;
Fig. 4: the schematic diagram that the pixel point is classified;
Fig. 5: two characteristic areas that mark off after the picture signal that optical mechaical scanning is obtained is handled i.e. zone 4 and zone 5;
Fig. 6: the image of Sheng Chenging after treatment
The image of a time t=0.05 second
The image of b time t=0.10 second
The image of c time t=0.15 second
The image of d time t=0.20 second
Embodiment
Read the detection target information one by one with ray machine scanning calorimeter imager and obtain its picture signal, its arbitrary frame image signal comprises the temperature value and the corresponding acquisition time thereof of all pixels, picture signal is stored in the known memory frame by frame, the display unit of optical mechaical scanning imaging system is converted to image with the picture signal in the known memory then, and picture signal needed to handle according to following steps successively before delivering to the display unit demonstration:
With reference to Fig. 1,
(1) reads in the temperature value and the corresponding acquisition time thereof of all pixels on a certain frame;
(2) adopt general wicket weighting smooth filtering method that the picture signal that detects target is carried out noise reduction process, concrete grammar is the value that replaces template center's pixel point with the mean value in the template, and it is big that coefficients should satisfy the horizontal direction weight factor, and vertically weight factor is little;
(3) temperature value and its time corresponding to all the pixel points in the image adopts method in common to carry out process of fitting treatment, obtain a temperature-time match curve, shown in the curve among Fig. 32, curve 1 has shown the distribution situation of all pixel point actual values in the image among Fig. 3.As can be seen from the figure, the value of each pixel point is lived apart in the both sides of matched curve 2, and the value of pixel point has illustrated that from the distance of matched curve this pixel point meets the degree of matched curve Changing Pattern, utilizes this point, by setting maximum allowable range d
mValue is carried out area dividing to the distance of matched curve with each pixel point according to each pixel value.As shown in Figure 4, this figure has illustrated the partition process of characteristic area.D in the present embodiment
mValue is 0.2 ℃, and curve 31 and curve 32 equate to the distance of matched curve 2 here, all are d
m, like this, whole coordinates regional has been divided into three parts, and topmost the point of a part is second class point, and mid portion is a first kind point, and nethermost part is the 3rd a class point.With reference to Fig. 4, in this processing procedure, there is not second class point, two characteristic areas that have only the first kind and the 3rd class point to constitute so far, are finished the characteristic area division first time in whole zone;
(4) respectively first kind point and the 3rd class point are carried out once more the processing procedure of step (3), the result, the classification results of this 2 class point all is to have only first kind point, so far, no longer it is further divided, the characteristic area in whole zone is divided and is finished;
(5) the pixel value of these two characteristic areas is carried out match, provide fitting coefficient, here three progression matches have been used, the fitting coefficient vector that obtains two characteristic areas is respectively (114.6734 123.6648-49.5705 41.0210) and (78.2266 82.2341-34.1242 38.7909), and promptly the value change with time of two characteristic areas is respectively:
f
1＝-114.6734+123.6648*t-49.5705*t
2+41.0210*t
3
f
2＝-78.2266+82.2341*t-34.1242*t
2+38.7909*t
3
(6) point set with two characteristic areas reverts on the former figure, as shown in Figure 5, as can be seen, utilizes fitting method to carry out characteristic area and divides, and has found the defective place exactly;
(7) owing to known the fitting parameter vector of two characteristic areas, the substitution time, just the digital picture of any time can have been obtained, shown in a figure, b figure, c figure and d figure among Fig. 6.
Like this, just the image that the optical mechaical scanning imaging system is gathered can be generated the image of no temporal resolution restriction.
Method according to embodiment 1, when the more weak samples of heat diffusion capabilities such as composite material, organic material and fiberglass being carried out the infrared thermal wave Non-Destructive Testing with opticmechanical scanning thermal imager, thermal change is not too fast, but because acquisition time is long, thermal diffusion also enlarges gradually to the influence of image data.In processing procedure, d
mValue generally can set greatly slightly, just can be generated as the image of optical mechaical scanning imaging system collection the image of no temporal resolution restriction.
Embodiment 3
In the infrared thermal wave Non-Destructive Testing to metal sample, it is very fast that heat wave changes, and the thermal change process concentrates on a certain frame or two frames of optical mechaical scanning image substantially, and the heat wave image that Direct observation is gathered can't be distinguished the position and the size of zones of different fully.By pixel being carried out the correlation Fitting Analysis first time, thermal process and thermal equilibrium state can be separated substantially, and then carry out the Fitting Analysis second time, just the different characteristic zone can be made a distinction.In processing procedure, d
mValue generally can set lessly because the heat diffusion capabilities of metal is extremely strong, zones of different influences each other, this can make the difference of zones of different become very not obvious, if d
mValue bigger, it can't be separated at all, certainly, d
mValue the influence that too small meeting makes noise handle data be set enlarge, should reduce system noise in the experiment as far as possible.
Embodiment 4
Method according to embodiment 3, in infrared thermal wave Non-Destructive Testing to blemish, the heat wave change procedure is equally very fast, by processing means such as matches pixel is carried out area dividing, can be generated as the high-frequency digital image to the image of optical mechaical scanning imaging system collection equally.But, in processing procedure, d
mValue should be according to the heat diffusion properties of actual sample, and the characteristics such as thickness on surface are comprehensively judged.
The extraction of characteristics of image when the method for area dividing can be used for the normal image processing among the embodiment 1.
The present invention finds by a series of experiment, pass through fitting method, pixel is carried out piecemeal, made full use of the advantage of optical mechaical scanning imaging system institute images acquired, reduced simultaneously of the influence of optical mechaical scanning imaging system self shortcoming again greatly, thereby the image of optical mechaical scanning imaging system collection has been generated as the image of no temporal resolution restriction its image data.D among the present invention
mBe that the maximum allowable range that criterion is set is the parameter of a key, it has great influence to the number of regions and the effect of dividing.By judgement, determine the d an of the best to actual conditions
mAfter, just can successfully finish the area dividing of pixel, calculate each regional fitting parameter vector then, just can obtain the numerical value of any time pixel after the substitution time.
Claims (5)
1. one kind generates the method for image without time resolution limit based on the optical mechaical scanning signal, comprising:
The optical mechaical scanning imaging system reads the detection target information one by one and obtains its picture signal, and picture signal is stored in the known memory frame by frame, and wherein, the information of each pixel point comprises temperature value and corresponding acquisition time thereof;
The display unit of optical mechaical scanning imaging system is converted to image with the picture signal in the known memory;
It is characterized in that before the picture signal in the known memory being delivered to the display unit demonstration, also will passing through following treatment step successively:
(1) reads in the temperature value and the time corresponding value thereof of all pixels on a certain frame;
(2) picture signal is carried out the noise reduction sonication;
(3) all pixels in the image are carried out correlation analysis, obtain the variation rule curve of all pixels on a certain frame, the i.e. temperature-time curve of match by general approximating method;
(4) calculate each pixel value to matched curve apart from d, according to each pixel value to apart from d each pixel point being carried out characteristic area and divides of matched curve, according to the maximum allowable range d of setting
mThe different situations of value and d magnitude relationship are divided into three classes with all pixel points:
The first kind: | d|＜d
m, these points belong to same characteristic area;
Second class: d〉d
m, forward step (3) to, the pixel point in the class is carried out process of fitting treatment once more, obtain a new temperature-time curve, then to the processing of pixel point repeating step (4) in the class;
The 3rd class: d＜-d
m, forward step (3) to, the pixel point in the class is carried out process of fitting treatment once more, obtain a new temperature-time curve, then to the processing of pixel point repeating step (4) in the class;
Till the classification results that obtains all was first kind pixel point, so far, the characteristic area partition process finished;
(5) all pixels in each characteristic area are carried out process of fitting treatment, the pixel that is in the same area has identical Changing Pattern, as long as promptly obtain a fitting parameter vector, and just can this regional variation characteristic of unique description;
(6) according to the fitting parameter vector of each characteristic area, substitution time t promptly can calculate the temperature value of any all pixels of the moment in the pairing time period of this two field picture;
At last the data image signal that generates is delivered to the display unit demonstration and obtained digital picture.
2. the method for generation image without time resolution limit according to claim 1, it is characterized in that to the concrete grammar that original image signal carries out noise reduction process be the value that replaces template center's pixel point with the mean value in the template, and it is big that coefficients should satisfy the horizontal direction weight factor, and vertically weight factor is little.
3. the method for generation image without time resolution limit according to claim 2 is characterized in that the pixel in each characteristic area obtains a fitting parameter vector a=(a by curve fit
0, a
1... a
N-1), promptly obtain the match change curve of all pixels in certain characteristic area
This curve is the variation rule curve of all pixels in this characteristic area.
4. the method for generation image without time resolution limit according to claim 3 is characterized in that the criterion d that sets
mFor the pixel value to the maximum allowable range between the curve, d
mSet point should be less than the greatest measure difference of zones of different greater than the maximum noise of system.
5. the method for generation image without time resolution limit according to claim 4 is characterized in that the number of fitting parameter vector equates with the number of characteristic area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100840424A CN100484191C (en) | 2004-10-18 | 2004-10-18 | Method for generating image without time resolution limit based on light machine scanning signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2004100840424A CN100484191C (en) | 2004-10-18 | 2004-10-18 | Method for generating image without time resolution limit based on light machine scanning signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1588982A CN1588982A (en) | 2005-03-02 |
| CN100484191C true CN100484191C (en) | 2009-04-29 |
Family
ID=34605152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2004100840424A Expired - Fee Related CN100484191C (en) | 2004-10-18 | 2004-10-18 | Method for generating image without time resolution limit based on light machine scanning signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100484191C (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008211717A (en) * | 2007-02-28 | 2008-09-11 | Brother Ind Ltd | Communication apparatus |
-
2004
- 2004-10-18 CN CNB2004100840424A patent/CN100484191C/en not_active Expired - Fee Related
Non-Patent Citations (7)
| Title |
|---|
| 利用图像的相关性实现图像抗信源丢失的编码. 宋建新，毕厚杰.南京邮电学院学报，第18卷第4期. 1998 |
| 利用图像的相关性实现图像抗信源丢失的编码. 宋建新，毕厚杰.南京邮电学院学报，第18卷第4期. 1998 * |
| 图像工程 上册：图象处理和分析. 章毓晋，194-200，清华大学出版社. 1999 无损检测技术及其应用. 张俊哲等，450-453，科学出版社. 1993 |
| 图像工程 上册：图象处理和分析. 章毓晋，194-200，清华大学出版社. 1999 * |
| 图象处理在脉冲加热红外热成像定量检测中的应用. 杨慧君，丁铁英，薛书文，宗明成.北方交通大学学报，第23卷第2期. 1999 |
| 图象处理在脉冲加热红外热成像定量检测中的应用. 杨慧君，丁铁英，薛书文，宗明成.北方交通大学学报，第23卷第2期. 1999 * |
| 无损检测技术及其应用. 张俊哲等，450-453，科学出版社. 1993 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1588982A (en) | 2005-03-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110389153B (en) | Hot defect detection method and hot defect detection system | |
| US7158674B2 (en) | Scene change detection apparatus | |
| CN101923711B (en) | SAR (Synthetic Aperture Radar) image change detection method based on neighborhood similarity and mask enhancement | |
| US7734064B2 (en) | Method and apparatus for classifying geological materials using image processing techniques | |
| Alzahrani et al. | A real-time edge detector: algorithm and VLSI architecture | |
| CN101655914B (en) | Training device, training method and detection method | |
| CN111652098B (en) | Product surface defect detection method and device | |
| KR102484094B1 (en) | Method and system for selecting recipes for defect inspection | |
| CN101459765B (en) | Method and circuit for correcting dead pixel in video signal | |
| CN109194954B (en) | Method, device and equipment for testing performance parameters of fisheye camera and storable medium | |
| US8873839B2 (en) | Apparatus of learning recognition dictionary, and method of learning recognition dictionary | |
| CN103389310A (en) | Online sub-pixel optical component damage detection method based on radiation calibration | |
| CN103561274B (en) | Video time domain tamper detection method for removing moving object shot by static camera lens | |
| CN113688830B (en) | Deep learning target detection method based on center point regression | |
| CN100433047C (en) | Device and method for detecting blurring of image | |
| CN100484191C (en) | Method for generating image without time resolution limit based on light machine scanning signal | |
| JP4626418B2 (en) | Object detection device and object detection method | |
| US20020191831A1 (en) | System and process for analyzing surface defects | |
| CN115546129A (en) | FPGA-based infrared image bad element detection method | |
| CN111640117B (en) | Method for searching leakage source position of building | |
| Xu et al. | A deep neural network-based intelligent detection model for manufacturing defects of automobile parts | |
| JP2004199391A (en) | Method and device for deciding threshold in image analysis, binarizing device and image analyzing device and information processing method with learning function and image analyzing device with learning function and, recording medium for these methods | |
| CN108491852B (en) | Blind pixel detection method for area array infrared focal plane | |
| JPH0620054A (en) | Method and apparatus for decision of pattern data | |
| JPH08214159A (en) | Image processing method and image processing unit |
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: 20090429 Termination date: 20151018 |
|
| EXPY | Termination of patent right or utility model |