CN109360189A - The method for detecting uncooled ir machine core image pixel defect point - Google Patents
The method for detecting uncooled ir machine core image pixel defect point Download PDFInfo
- Publication number
- CN109360189A CN109360189A CN201811084548.3A CN201811084548A CN109360189A CN 109360189 A CN109360189 A CN 109360189A CN 201811084548 A CN201811084548 A CN 201811084548A CN 109360189 A CN109360189 A CN 109360189A
- Authority
- CN
- China
- Prior art keywords
- pixel
- image
- machine core
- defect point
- average
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Abstract
A kind of method for detecting uncooled ir machine core image pixel defect point of the present invention, it include: that N frame machine core image is acquired by image pick-up card, the N frame machine core image for reading acquisition, analyzes to obtain the grayscale information of each frame machine core image the gray value of N frame machine core image；Take average calculating to obtain the average image N frame machine core image respective pixel gray value；Calculate pixel coordinate (x in the average image, y) the judgment threshold BP of contrast district median difference MedDif [x] [y] and machine core current state, the pixel is judged as bright defect point if MedDif [x] [y] > BP, the pixel is judged as dark defect point if MedDif [x] [y] <-BP, and otherwise the pixel is judged as not being picture element flaw point.The present invention can it is objective, accurate, rapidly identify defect point on image, can accurately judge that pixel size shared by each defect point and the position coordinates where defect point.
Description
Technical field
The present invention relates to uncooled ir machine core image pixel defect point (show as in uncooled ir machine core image with
The different brighter or darker pixel of the numerous pixels of surrounding, abbreviation defect point) technical field, it is based on more particularly to one kind
The method of image pick-up card detection uncooled ir machine core image pixel defect point.
Background technique
The picture quality of machine core is the important indicator of user's concern, and the quantity and size of image deflects point are exactly evaluation figure
One of the important indicator of image quality amount.
Original detection method to infrared machine core image deflects point is eye recognition, by the way that the parameters of machine core are adjusted
It is extremely moderate, a homogeneous background is provided for machine core, eye recognition goes out the number of defect point and the about shared picture of defect point on image
Plain size, and record the general area where defect point.Eye recognition defect point number is feasible, but is difficult to accomplish accurately
It identifies the position coordinates at pixel size shared by each defect point and place, can only obtain a general judgement by rule of thumb." people
Outlook is other " there is certain subjectivity, and also detection precision is inadequate, and detection efficiency is lower.
Summary of the invention
The present invention is directed to the status of current uncooled ir machine core image deflects point detection, provides a kind of based on Image Acquisition
The method of card detection uncooled ir machine core image pixel defect point, can it is objective, accurate, rapidly identify defect on image
Point, and can accurately judge that pixel size shared by each defect point and the position coordinates where defect point.
The present invention is to solve above-mentioned technical problem by following technical proposals:
The present invention provides a kind of method for detecting uncooled ir machine core image pixel defect point, it is characterized in that, packet
Include following steps:
S1, N frame machine core image is acquired by image pick-up card, the N frame machine core image of acquisition is read, to the N frame machine core figure
The gray value of picture is analyzed to obtain the grayscale information of each frame machine core image, and N is positive integer；
S2, take average calculating to obtain a average image N frame machine core image respective pixel gray value；
S3, machine core image is judged by analysis the average image with the presence or absence of defect point, calculate pixel in the average image and sit
The judgment threshold BP for marking contrast district median difference MedDif [x] [y] and machine core current state of (x, y), if MedDif [x] [y] >
Then the pixel is judged as bright defect point to BP, and the pixel is judged as dark defect point if MedDif [x] [y] <-BP, otherwise should
Pixel is judged as not being picture element flaw point.
Preferably, in step s3, when judging whether coordinate belongs to defect point for the pixel of (x, y), in conjunction with pixel week
The gray value of 15*15 pixel region is enclosed to judge, which is known as the contrast district for the pixel that coordinate is (x, y)；
Utilize formulaCalculating coordinate in the average image is (x, y) pixel
Gray value, wherein Gray [n] [x] [y] is the gray value that coordinate is (x, y) pixel in n-th frame image, and VGray [x] [y] is
Coordinate is the gray value of (x, y) pixel, n≤N in the average image；
By in the average image in the contrast district of pixel coordinate (x, y) all pixels gray value according to sorting from small to large, and
Most intermediate grey scale pixel value will be arranged in and be known as contrast district intermediate value Med [x] [y]；
Contrast district median difference is calculated using formula MedDif [x] [y]=VGray [x] [y]-Med [x] [y].
Preferably, in step s3, in the average image in the contrast district of pixel coordinate (x, y) all pixels gray value it is flat
Mean value is known as comparing plot mean Aver [x] [y], all pixels gray scale in the contrast district of pixel coordinate (x, y) in the average image
The standard deviation of value is defined as contrast district uniformity Un [x] [y], and the corresponding contrast district uniformity of all pixels is flat in the average image
Mean value is known as the average Uniformity Un_AVER of image:
Wherein: the number of xmax representative image horizontal pixel, the number of ymax representative image vertical pixel；
Judgment threshold is calculated using formula BP=8.196*Un_AVER.
On the basis of common knowledge of the art, above-mentioned each optimum condition, can any combination to get each preferable reality of the present invention
Example.
The positive effect of the present invention is that:
The present invention relates to the detection method of uncooled ir machine core image deflects point, in conjunction with image/video acquisition card technique,
The methods of computer programming, data statistical analysis method and mathematical modeling develop it is a kind of can it is objective, quick, accurate,
The method for automatically detecting that machine core image deflects point.Present method solves the precise positionings of defect point position coordinates, defect point
The problems such as statistics of pixel size shared by the clearly division of region, continuous defect point.
Detailed description of the invention
Fig. 1 is the process of the method for the detection uncooled ir machine core image pixel defect point of present pre-ferred embodiments
Figure.
Fig. 2 is that the coordinate of present pre-ferred embodiments is the contrast district figure of (x, y) pixel.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, and for explaining only the invention, and is not considered as limiting the invention.On the contrary, this
The embodiment of invention includes all changes fallen within the scope of the spiritual and intension of attached claims, modification and is equal
Object.
As shown in Figure 1, the present embodiment provides a kind of method for detecting uncooled ir machine core image pixel defect point, packet
Include following steps:
Step 101 acquires N frame machine core image by image pick-up card, the N frame machine core image of acquisition is read, to the N frame machine
The gray value of core image is analyzed to obtain the grayscale information of each frame machine core image, and N is positive integer；
Step 102 takes average calculating to obtain a average image N frame machine core image respective pixel gray value；
Step 103 judges that machine core image with the presence or absence of defect point, calculates picture in the average image by analysis the average image
The judgment threshold BP of contrast district median difference MedDif [x] [y] and machine core current state of plain coordinate (x, y), if MedDif [x]
Then the pixel is judged as bright defect point to [y] > BP, and the pixel is judged as dark defect point if MedDif [x] [y] <-BP,
Otherwise the pixel is judged as not being picture element flaw point.
Lower mask body introduction judges machine core image with the presence or absence of defect point by analysis the average image information.
It, should be in conjunction with 15*15 pixel region around the pixel when judging whether coordinate belongs to defect point for the pixel of (x, y)
The gray value in domain judges, as shown in Fig. 2, the region is known as the contrast district that coordinate is (x, y) pixel, abbreviation contrast district.
One, contrast district median difference (MedDif [x] [y]) derives calculation method
Defining Gray [n] [x] [y] is the gray value that coordinate is (x, y) pixel in n-th frame image, and VGray [x] [y] is flat
Coordinate is the gray value of (x, y) pixel in equal image, if the video totalframes of acquisition is N, then:
Formula 1:
In the average image, all pixels gray value is from small to large in the contrast district of pixel coordinate (x, y) in the average image
Sequence, is arranged in most intermediate grey scale pixel value and is known as contrast district intermediate value, it represents all pixels tonal gradation in contrast district
Mean level, indicated with Med [x] [y], the difference of VGray [x] [y] and Med [x] [y] are known as contrast district median difference, its generation
The table level of difference of the grey scale pixel value and its surrounding pixel gray scale is indicated with MedDif [x] [y]:
Formula 2:
Med [x] [y]=Median VGray [x+n] [y+m] | and n, m=-7, -6, -5, -4, -3, -2, -1,0,1,2,3,
4,5,6,7}
Formula 3:MedDif [x] [y]=VGray [x] [y]-Med [x] [y]
If coordinate is that the pixel of (x, y) is not defect point, contrast district median difference MedDif [x] [y] absolute value is very
It is small, 0 is leveled off to, conversely, the point belongs to bright defect point when MedDif [x] [y] is greater than BP, when MedDif [x] [y] is less than-BP
When, which belongs to dark defect point, and BP is referred to as judgment threshold.
Two, the determination method of judgment threshold (BP)
In the average image, the average value of all pixels gray scale claims in the contrast district of pixel coordinate (x, y) in the average image
It to compare plot mean, is indicated with Aver [x] [y], all pixels gray scale in the contrast district of pixel coordinate (x, y) in the average image
Standard deviation be defined as contrast district uniformity, it represents the dispersion degree in contrast district between all pixels gray scale, with Un [x]
[y] is indicated:
Formula 4:
Formula 5:
In whole secondary the average image, the average value of the corresponding contrast district uniformity of all pixels is known as the average homogeneous of image
Property, abbreviation average Uniformity, it represents the uniformity coefficient between whole sub-picture pixel grey scale, it is indicated with Un_AVER:
Formula 6:
Wherein: the number of xmax representative image horizontal pixel；The number of ymax representative image vertical pixel.
In order to determine bright dark threshold value BP, following test, following Tables 1 and 2 have been done using the machine core of known existing defects point
Middle is the test situation of model 384*288 Yu 640*512 machine core respectively." state " in table refers to same branch machine core in difference
State under contrast, " defect point median difference detail drawing " are the enlarged drawings of defect point, and the number in figure is in respective coordinates point
Value difference.
Following rule is found by the data of Tables 1 and 2: first, normal pixel in defect pixel point and contrast district
Gray difference it is bigger, the corresponding median difference absolute value of the pixel is bigger, on the contrary just it is smaller；Second, for same branch machine core,
When average Uniformity Un_AVER is bigger, the median difference of the same defect pixel point on machine core image is also bigger, this illustrates Un_
The median difference of AVER and defect pixel point, which exists, to be positively correlated；Third, a defect point are made of one or more defect pixels, are lacked
The center gray difference of trapping spot is maximum, and from center, pixel grey scale difference outward gradually becomes smaller.
Table 1
Table 1
By upper Tables 1 and 2 it is found that a defect point is usually made of two or more picture element flaw point.It is fixed
Minimum median difference in an adopted defect point is MF_MIN, in order to prove judgment threshold BP and average Uniformity Un_AVER positive
It closes, has done further test using the known machine core with defect point, test result is shown in Table 3 and table 4.
Coefficient MIN_AK in table 3 and table 4 is the ratio of MF_MIN and Un_AVER, it may be assumed that
Formula 7:MIN_AK=MF_MIN/Un_AVER；
For numbering the machine core for being MDIC052-154, under three state, the BP of gain hour should be less than 9, when in gain
BP should be less than 13, BP when gain is big should be less than 16, and software detection result could be consistent with human eye detection, and BP is with gain
Become larger and increase, it is seen that BP is not a definite value.By table 3 and table 4, we are also found that when becoming larger with machine core gain,
Average Uniformity Un_AVER is also being increased with it, this illustrates judgment threshold BP and Un_AVER, and there is inevitable to contact.
Assuming that formula 8:BP=Un_AVER*AK
Wherein AK is the positive correlation coefficient of judgment threshold and average Uniformity.
Gained formula 9:MF_MIN=Un_AVER*MIN_AK is derived by formula 7
In order to avoid the missing inspection of defect point, BP must satisfy the relationship of BP < MF_MIN, and can by formula 8 and the derivation of formula 9
: formula 10:AK < MIN_AK
Table 3
Table 4
Meanwhile a large number of experiments has also been made using the machine core of known intact trapping spot, see Table 5 for details and table 6 for test situation, in table
MF_MAX indicates the maximum value of all pixels contrast district median difference in machine core the average image, and coefficient MAX_AK is MF_MAX and Un_
The ratio of AVER, it may be assumed that
Formula 11:MF_MAX=MAX MedDif [x] [y] | x=1:2:xmax, y=1:2:ymax }
Formula 12:MAX_AK=MF_MAX/Un_AVER
By test data it is found that the erroneous judgement of defect point will not just occur only when BP meets the relationship of BP > MF_MAX.
Therefore, it can be obtained by formula 8 and the derivation of formula 12:
Formula 13:AK > MAX_AK
Table 5
Table 6
Maximum MAX_AK=7.652, the smallest MIN_AK=known in the experimental data of table 3, table 4, table 5 and table 6
8.739.MAX_AK < AK < MIN_AK can be obtained by formula 10 and formula 13, it may be assumed that 7.652 < AK < 8.739, because 7.652 < AK <
8.739 be true statement, so the hypothesis to formula 8 is set up.
According to a large amount of verification experimental verification, AK takes the median of maximum MAX_AK and minimum MIN_AK relatively reasonable, i.e. AK=
8.196, BP=8.196*Un_AVER can be obtained according to formula 8.
According to above-mentioned described method, other types or other model machine cores can determine AK's with same method
Value, and then obtain the value of judgment threshold BP.
Non-brake method is realized in conjunction with video acquisition card technique by computer software programming according to the algorithm in technical solution
Infrared machine core image deflects point is objective, it is accurate, rapidly detect.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is by appended claims and its equivalent limits.
Claims (3)
1. it is a kind of detect uncooled ir machine core image pixel defect point method, which is characterized in that itself the following steps are included:
S1, N frame machine core image is acquired by image pick-up card, the N frame machine core image of acquisition is read, to the N frame machine core image
Gray value is analyzed to obtain the grayscale information of each frame machine core image, and N is positive integer；
S2, take average calculating to obtain a average image N frame machine core image respective pixel gray value；
S3, machine core image is judged by analysis the average image with the presence or absence of defect point, calculate pixel coordinate in the average image (x,
Y) the judgment threshold BP of contrast district median difference MedDif [x] [y] and machine core current state, should if MedDif [x] [y] > BP
Pixel is judged as bright defect point, and the pixel is judged as dark defect point if MedDif [x] [y] <-BP, otherwise the pixel quilt
It is judged to not being picture element flaw point.
2. the method for detection uncooled ir machine core image pixel defect point according to claim 1, which is characterized in that
In step S3, when judging whether coordinate belongs to defect point for the pixel of (x, y), in conjunction with 15*15 pixel region around the pixel
Gray value judge that the region is known as the contrast district for the pixel that coordinate is (x, y)；
Utilize formulaCalculate the ash that coordinate in the average image is (x, y) pixel
Angle value, wherein Gray [n] [x] [y] is the gray value that coordinate is (x, y) pixel in n-th frame image, and VGray [x] [y] is average
Coordinate is the gray value of (x, y) pixel, n≤N in image；
By all pixels gray value, and will row according to sorting from small to large in the contrast district of pixel coordinate (x, y) in the average image
It is listed in most intermediate grey scale pixel value and is known as contrast district intermediate value Med [x] [y]；
Contrast district median difference is calculated using formula MedDif [x] [y]=VGray [x] [y]-Med [x] [y].
3. the method for detection uncooled ir machine core image pixel defect point according to claim 2, which is characterized in that
In step S3, to be known as contrast district flat for the average value of all pixels gray value in the contrast district of pixel coordinate (x, y) in the average image
Mean value Aver [x] [y], the standard deviation of all pixels gray value is defined as in the contrast district of pixel coordinate (x, y) in the average image
Contrast district uniformity Un [x] [y], the average value of the corresponding contrast district uniformity of all pixels is known as the flat of image in the average image
Equal uniformity Un_AVER:
Wherein: the number of xmax representative image horizontal pixel, the number of ymax representative image vertical pixel；
Judgment threshold is calculated using formula BP=8.196*Un_AVER.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811084548.3A CN109360189B (en) | 2018-09-18 | 2018-09-18 | Method for detecting pixel defect point of image of uncooled infrared movement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811084548.3A CN109360189B (en) | 2018-09-18 | 2018-09-18 | Method for detecting pixel defect point of image of uncooled infrared movement |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109360189A true CN109360189A (en) | 2019-02-19 |
| CN109360189B CN109360189B (en) | 2021-08-10 |
Family
ID=65350970
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811084548.3A Active CN109360189B (en) | 2018-09-18 | 2018-09-18 | Method for detecting pixel defect point of image of uncooled infrared movement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109360189B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109916914A (en) * | 2019-04-10 | 2019-06-21 | 清华大学深圳研究生院 | A kind of product defects detection method and device |
| CN111839180A (en) * | 2020-07-07 | 2020-10-30 | 胡飞青 | Intelligent wheel-by-wheel operation opportunity identification platform |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750701A (en) * | 2012-06-15 | 2012-10-24 | 西安电子科技大学 | Method for detecting spissatus and spissatus shadow based on Landsat thematic mapper (TM) images and Landsat enhanced thematic mapper (ETM) images |
| CN105787906A (en) * | 2016-03-25 | 2016-07-20 | 北京环境特性研究所 | Method and system for rejecting bright noises from infrared image |
| US20170063656A1 (en) * | 2015-08-31 | 2017-03-02 | Viasat, Inc. | Methods and systems for transportation vessel network communication service anomaly detection |
| CN107486415A (en) * | 2017-08-09 | 2017-12-19 | 中国计量大学 | Thin bamboo strip defect on-line detecting system and detection method based on machine vision |
-
2018
- 2018-09-18 CN CN201811084548.3A patent/CN109360189B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102750701A (en) * | 2012-06-15 | 2012-10-24 | 西安电子科技大学 | Method for detecting spissatus and spissatus shadow based on Landsat thematic mapper (TM) images and Landsat enhanced thematic mapper (ETM) images |
| US20170063656A1 (en) * | 2015-08-31 | 2017-03-02 | Viasat, Inc. | Methods and systems for transportation vessel network communication service anomaly detection |
| CN105787906A (en) * | 2016-03-25 | 2016-07-20 | 北京环境特性研究所 | Method and system for rejecting bright noises from infrared image |
| CN107486415A (en) * | 2017-08-09 | 2017-12-19 | 中国计量大学 | Thin bamboo strip defect on-line detecting system and detection method based on machine vision |
Non-Patent Citations (1)
| Title |
|---|
| 李成立: ""红外制导武器探测器成像检测及补偿仿真"", 《计算机仿真》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109916914A (en) * | 2019-04-10 | 2019-06-21 | 清华大学深圳研究生院 | A kind of product defects detection method and device |
| CN109916914B (en) * | 2019-04-10 | 2021-07-02 | 清华大学深圳研究生院 | Product defect detection method and device |
| CN111839180A (en) * | 2020-07-07 | 2020-10-30 | 胡飞青 | Intelligent wheel-by-wheel operation opportunity identification platform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109360189B (en) | 2021-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6777726B2 (en) | Colony contrast collection | |
| CN106412573B (en) | A kind of method and apparatus of detector lens stain | |
| CN1993708B (en) | Image processing apparatus and method, image sensing apparatus | |
| CN107784651B (en) | Fuzzy image quality evaluation method based on fuzzy detection weighting | |
| CN110490914A (en) | It is a kind of based on brightness adaptively and conspicuousness detect image interfusion method | |
| CN111754906B (en) | Method for detecting brightness uniformity of curved surface display screen | |
| CN108827969A (en) | Metal parts surface defects detection and recognition methods and device | |
| CN110264426A (en) | Image distortion correction method and apparatus | |
| CN109152535A (en) | Skin diagnosis device and skin diagnosis method | |
| CN109886960A (en) | The method of glass edge defects detection based on machine vision | |
| CN108876756A (en) | The measure and device of image similarity | |
| CN106067177A (en) | HDR scene method for detecting and device | |
| TWI376648B (en) | Method and device for keeping image background by multiple gauss models | |
| CN109360189A (en) | The method for detecting uncooled ir machine core image pixel defect point | |
| JP4060558B2 (en) | Defect inspection method and apparatus | |
| CN111062961A (en) | Contact lens edge defect detection method based on deep learning | |
| CN108844957A (en) | A kind of tobacco leaf oil new detecting method based on machine vision | |
| CN106878636B (en) | Defect pixel point detecting method, device and picture processing chip | |
| JPWO2014013792A1 (en) | Noise evaluation method, image processing apparatus, imaging apparatus, and program | |
| CN108965749A (en) | Defect pixel detection and means for correcting and method based on texture recognition | |
| CN1323545C (en) | Method for determining automatic detection threshold of bad pixel of medical image | |
| CN110674737B (en) | Iris recognition enhancement method | |
| CN113313638A (en) | Appearance defect detection method | |
| CN117011291A (en) | Watch shell quality visual detection method | |
| JP2022514427A (en) | Systems and methods for monitoring bacterial growth of bacterial colonies and predicting colony biomass |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |