CN114785965B - Automatic hyperspectral image exposure method and system based on COPOD algorithm - Google Patents
Automatic hyperspectral image exposure method and system based on COPOD algorithm Download PDFInfo
- Publication number
- CN114785965B CN114785965B CN202210415974.0A CN202210415974A CN114785965B CN 114785965 B CN114785965 B CN 114785965B CN 202210415974 A CN202210415974 A CN 202210415974A CN 114785965 B CN114785965 B CN 114785965B
- Authority
- CN
- China
- Prior art keywords
- hyperspectral
- image
- probability
- representing
- calculating
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000001186 cumulative effect Effects 0.000 claims abstract description 34
- 241000039077 Copula Species 0.000 claims abstract description 29
- 230000002146 bilateral effect Effects 0.000 claims abstract description 27
- 238000004458 analytical method Methods 0.000 claims abstract description 26
- 239000013598 vector Substances 0.000 claims abstract description 25
- 238000005315 distribution function Methods 0.000 claims abstract description 20
- 230000002159 abnormal effect Effects 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims description 38
- 230000008569 process Effects 0.000 claims description 21
- 238000001228 spectrum Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 33
- 238000001514 detection method Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
- H04N23/73—Circuitry for compensating brightness variation in the scene by influencing the exposure time
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Abstract
The application discloses a hyperspectral image automatic exposure method and a hyperspectral image automatic exposure system based on a COCOPOD algorithm, which relate to the technical field of image preprocessing and have the technical scheme that: acquiring a hyperspectral original image with exposure time; calculating an image feature vector set in each hyperspectral original image by taking the K moment as a main component; calculating a bilateral experience cumulative distribution function based on the image feature vector set; calculating an empirical Copula function based on the bilateral empirical cumulative distribution function; estimating bilateral tail probability values of joint distribution in all dimensions through an empirical Copula function; obtaining an analysis result of overexposure, darkness or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides; the exposure time is adjusted for the abnormal exposure condition until the exposure is normal. The application can ensure good accuracy and generalization by utilizing the feature vector of the original image and the experience coupler function.
Description
Technical Field
The application relates to the technical field of image preprocessing, in particular to a hyperspectral image automatic exposure method and system based on a COCOPOD algorithm.
Background
The hyperspectral camera can fully exert the high-resolution advantage of the hyperspectral camera in the spectrum dimension only under the condition of normal exposure. And because of the challenges of long imaging time, large memory of the output image and the like, the mode of manually adjusting the exposure time of the camera according to the sampling environment to optimize the imaging effect is lagged and complicated. The existing automatic exposure algorithm is mostly based on the traditional RGB camera and is difficult to apply to the hyperspectral camera. Therefore, how to research and design a hyperspectral image automatic exposure method and system based on COPOD algorithm, which can overcome the defects, is a problem which we need to solve at present.
Disclosure of Invention
In order to solve the defects in the prior art, the application aims to provide the hyperspectral image automatic exposure method and the hyperspectral image automatic exposure system based on the COCOPOD algorithm, which efficiently complete the automatic exposure function of the hyperspectral camera by utilizing the feature vectors corresponding to the original image under different exposure time, and can ensure good accuracy and generalization by utilizing the feature vectors of the original image and the experience coupler function.
The technical aim of the application is realized by the following technical scheme:
in a first aspect, there is provided a hyperspectral image automatic exposure method based on a COPOD algorithm, including the steps of:
acquiring a hyperspectral original image with exposure time;
calculating an image feature vector set in each hyperspectral original image by taking the K moment as a main component;
calculating a bilateral experience cumulative distribution function based on the image feature vector set;
calculating an empirical Copula function based on the bilateral empirical cumulative distribution function;
estimating bilateral tail probability values of joint distribution in all dimensions through an empirical Copula function;
obtaining an analysis result of overexposure, darkness or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides;
the exposure time is adjusted for the abnormal exposure condition until the exposure is normal.
Further, the process of obtaining the hyperspectral original image specifically includes:
collecting an original hyperspectral image with exposure time in a line scanning type or snapshot type by adopting a hyperspectral camera;
clipping the original hyperspectral image into an image with uniform height, width and spectrum channel number;
and naming the cut images in a unified format and recording corresponding exposure time to obtain hyperspectral original images.
Further, the calculating process of the image feature vector set specifically includes:
and calculating an index set corresponding to the pixel value statistic by using the K moment as a main band.
The pixel value statistics include: the first order origin moment, the second order center moment, the third order center distance, the fourth order center moment, the entropy value and the pixel proportion of the brightness of each channel pixel value are larger than a given value.
Further, the calculation formula of the entropy value specifically includes:
wherein X represents a random variable of pixel gray values of a certain frame of the hyperspectral image; entopy (X) represents the entropy value of the random variable X; x is X i Representing the actual value of the ith sample; p (x=x) i ) Indicating that the sample has a value of X i Is a frequency of (2); n represents the total number of samples; n represents the selected sample size.
Further, the pixel value statistic further comprises a combination statistic formed by at least two of a first order origin moment, a second order center moment, a third order center distance, a fourth order center moment, an entropy value and a pixel proportion with brightness larger than a given value.
Further, the calculation process of the bilateral experience cumulative distribution function specifically comprises the following steps:
the left cumulative probability density function is calculated, and the calculation formula is specifically as follows:
the right cumulative probability density function is calculated, and the calculation formula is specifically as follows:
calculating a skewness coefficient, wherein a calculation formula specifically comprises:
wherein ,a left cumulative probability density function representing a d-th dimensional feature; />A right cumulative probability density function representing a d-th dimensional feature; b d A skewness factor representing a d-th dimensional feature; x represents the order statistic of the samples; x is X i Representing the actual value of the ith sample; />Representing a sample mean; n represents the selected sample size.
Further, the calculation process of the empirical Copula function specifically includes:
left-hand empirical Copula observations were calculated from hyperspectral image samples:
the right-hand empirical Copula observation was calculated from the hyperspectral image samples:
calculating Copula observation statistics: if b d < 0, thenOtherwise, go (L)>
wherein ,represent left experience values; />Represent the right experience value; />Representing Copula observations; x is x i A sequence statistic representing ordinal number i; b d A skewness factor representing a d-th dimensional feature; />Representing observed statistics.
Further, the calculation process of the probability value of the two-sided tail end specifically includes:
from the hyperspectral image samples, the left tail probability is calculated:
from the hyperspectral image samples, the right tail probability is calculated:
calculating the integral tail probability according to the hyperspectral image sample:
wherein ,pl Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability; d represents a dimension feature sequence number;left experience values representing the j-th dimensional feature; />Right experience values representing the j-th dimensional feature; representation ofObservation statistics for the j-th dimension feature.
Further, the analysis result obtaining process specifically includes:
obtaining a judgment index score according to analysis of the probability values of the two tail ends, wherein a calculation formula of the index score is specifically as follows:
wherein ,S(xi ) Representing a decision index score; p is p l Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability;
and repeatedly calculating to obtain the judgment index scores of all samples, and analyzing and judging overexposure, overdrising or normal exposure of the hyperspectral original image according to the distribution interval of the judgment index scores.
In a second aspect, there is provided a hyperspectral image automatic exposure system based on a COPOD algorithm, comprising:
the image acquisition module is used for acquiring a hyperspectral original image with exposure time;
the characteristic calculation module is used for calculating an image characteristic vector set in each hyperspectral original image by taking the K moment as a main component;
a first function module for calculating a bilateral empirical cumulative distribution function based on the set of image feature vectors;
the second function module is used for calculating an experience Copula function based on the bilateral experience cumulative distribution function;
the probability estimation module is used for estimating bilateral tail end probability values which are distributed in a combined mode in all dimensions through an empirical Copula function;
the result analysis module is used for obtaining the analysis result of overexposure, overdose or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides;
and the automatic adjustment module is used for adjusting the exposure time according to the abnormal exposure condition until the exposure is normal.
Compared with the prior art, the application has the following beneficial effects:
1. compared with the traditional automatic exposure of the visible light image, the automatic exposure method of the hyperspectral image based on the COCOPOD algorithm provided by the application has the characteristics that the hyperspectral image is numerous in image spectrum, large in data volume, obvious in imaging difference of different spectrum sections and the like; according to the application, the automatic exposure function of the hyperspectral camera is efficiently completed by utilizing the feature vectors corresponding to the original images under different exposure time, and good accuracy and generalization can be ensured by utilizing the feature vectors of the original images and the experience coupler function;
2. in the application process, the application can automatically adjust the exposure time by combining the returned result without adding additional detection equipment and detection equipment, realizes the automatic exposure function, and has strong adaptability and practicability to application scenes.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:
FIG. 1 is a flow chart in an embodiment of the application;
FIG. 2 is a probability density diagram in an embodiment of the application, a is bias, b is entropy, c is the pixel duty ratio of pixel values below 10, d is the pixel duty ratio of pixel values above 200, and e is the ratio of standard deviation to mean;
FIG. 3 is a probability density plot of an embodiment of the application, a being the mean, b being the standard deviation, c being the kurtosis value;
FIG. 4 is a graph of probability density for an embodiment of the application, a being a graph of the proportion of pixel values above 253, b being a graph of the proportion of pixel values below 4;
FIG. 5 is a schematic diagram of the detection result in the embodiment of the present application;
fig. 6 is a system block diagram in an embodiment of the application.
Detailed Description
For the purpose of making apparent the objects, technical solutions and advantages of the present application, the present application will be further described in detail with reference to the following examples and the accompanying drawings, wherein the exemplary embodiments of the present application and the descriptions thereof are for illustrating the present application only and are not to be construed as limiting the present application.
Example 1: the hyperspectral image automatic exposure method based on the COPOD algorithm, as shown in figure 1, comprises the following steps:
s1: acquiring a hyperspectral original image with exposure time;
s2: calculating an image feature vector set in each hyperspectral original image by taking the K moment as a main component;
s3: calculating a bilateral experience cumulative distribution function based on the image feature vector set;
s4: calculating an empirical Copula function based on the bilateral empirical cumulative distribution function;
s5: estimating bilateral tail probability values of joint distribution in all dimensions through an empirical Copula function;
s6: obtaining an analysis result of overexposure, darkness or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides;
s7: the exposure time is adjusted for the abnormal exposure condition until the exposure is normal.
It should be noted that, the hyperspectral original image can also be replaced by the multispectral original image, and the technology provided by the application is also applicable.
In step S1, the process of acquiring the hyperspectral raw image specifically includes: collecting an original hyperspectral image with exposure time in a line scanning type or snapshot type by adopting a hyperspectral camera; clipping the original hyperspectral image into an image with uniform height, width and spectrum channel number; and naming the cut images in a unified format and recording corresponding exposure time to obtain hyperspectral original images.
For example, an image cropped to h×w×c, where h×w×c is the height, width, and number of spectral channels, respectively, may be named after the image containing its exposure time field.
The calculation process of the image feature vector set specifically comprises the following steps: and calculating an index set corresponding to the pixel value statistic by using the K moment as a main band. The pixel value statistics include: the first-order origin moment, the second-order middle moment, the third-order center distance, the fourth-order center moment, the entropy value and the pixel proportion of the brightness of each channel pixel value are larger than given values.
Moment statistics are important statistics of random variable distribution functions, and regarding definition of K-order central moment of a sample distribution, the number of stages is as follows:
the integral form is as follows:
wherein X represents a random variable of pixel gray values of a certain frame of the hyperspectral image; x is X i The actual value of the sample, namely the specific pixel gray value of a certain frame of the hyperspectral image; f (x) is a probability density function of the pixel gray value distribution. The moment is used as an important and basic statistic for measuring the property of the distribution function, so that corresponding indexes are obtained by definition and calculation of the moment statistic, and the properties and characteristics of different types of images can be reflected.
The first order origin moment, also called the mean, is expressed as:
the second order central moment, also called variance, is expressed as:
the third order center distance, also called skewness, is expressed as:
the fourth order central moment, also known as kurtosis, is expressed as:
the calculation formula of the entropy value is specifically as follows:
wherein X represents a random variable of pixel gray values of a certain frame of the hyperspectral image; entopy (X) represents the entropy value of the random variable X; x is X i Representing the actual value of the ith sample; p (x=x) i ) Indicating that the sample has a value of X i Is a frequency of (2); n represents the total number of samples; n represents the selected sample size.
The pixel value statistic also comprises a combination statistic consisting of at least two of a first order origin moment, a second order center moment, a third order center distance, a fourth order center moment, an entropy value and a pixel proportion with brightness larger than a given value, and can also be a transformation statistic.
For example, a combined statistic of the mean divided by the standard deviation is more sensitive as an indicator. Standard deviation is the square root of the variance of a sample, defined as:the average value-standard deviation ratio is a combination index, and the expression is as follows: />Wherein μ above represents the mean value and δ represents the standard deviation.
The gray-scale too high pixel duty ratio is defined as the percentage of the gray value of a pixel of a certain frame in the hyperspectral image exceeding a certain threshold value, and the expression is as follows:wherein a is a threshold value.
Using the existing hyperspectral image sample, calculating the index corresponding to each statistic of step S2, and obtaining probability density diagrams of the calculated index as shown in fig. 2, 3 and 4.
The calculation process of the bilateral experience cumulative distribution function specifically comprises the following steps:
(1) The left cumulative probability density function is calculated, and the calculation formula is specifically as follows:
(2) The right cumulative probability density function is calculated, and the calculation formula is specifically as follows:
(3) Calculating a skewness coefficient, wherein a calculation formula specifically comprises:
wherein ,a left cumulative probability density function representing a d-th dimensional feature; />A right cumulative probability density function representing a d-th dimensional feature; bd represents the skewness coefficient of the d-th dimensional feature; x represents the order statistic of the samples; xi represents the actual value of the ith sample; />Representing a sample mean; n represents the selected sample size.
The calculation process of the empirical Copula function is specifically as follows:
(1) Left-hand empirical Copula observations were calculated from hyperspectral image samples:
(2) The right-hand empirical Copula observation was calculated from the hyperspectral image samples:
(3) Calculating Copula observation statistics: if bd<0, thenOtherwise, go (L)>
wherein ,represent left experience values; />Represent the right experience value; />Representing Copula observations; xi represents the order statistic for ordinal number i; bd represents the skewness coefficient of the d-th dimensional feature; />Representing observed statistics.
The calculation process of the probability value of the two side tail ends specifically comprises the following steps:
(1) From the hyperspectral image samples, the left tail probability is calculated:
(2) From the hyperspectral image samples, the right tail probability is calculated:
(3) Calculating the integral tail probability according to the hyperspectral image sample:
wherein ,pl Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability; d represents a dimension feature sequence number;left experience values representing the j-th dimensional feature; />Right experience values representing the j-th dimensional feature; representation ofObservation statistics for the j-th dimension feature.
The analysis result is obtained by the following steps:
(1) Obtaining a judgment index score according to analysis of the probability values of the two tail ends, wherein a calculation formula of the index score is specifically as follows:
wherein ,S(xi ) Representing a decision index score; p is p l Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability;
(2) And repeatedly calculating to obtain the judgment index scores of all samples, and analyzing and judging overexposure, overdrising or normal exposure of the hyperspectral original image according to the distribution interval of the judgment index scores.
The detection results obtained for the hyperspectral images at different exposure times using the COPOD anomaly detection algorithm are shown in fig. 5. Wherein the vertical axis value 1 represents abnormal exposure, and the value 2 is normal exposure; the horizontal axis represents exposure time, the left anomaly represents underexposed over-dark sample, and the right anomaly represents over-exposed sample.
Example 2: the hyperspectral image automatic exposure system based on the COPOD algorithm comprises an image acquisition module, a feature calculation module, a first function module, a second function module, a probability estimation module, a result analysis module and an automatic adjustment module as shown in fig. 6.
The image acquisition module is used for acquiring a hyperspectral original image with exposure time; the characteristic calculation module is used for calculating an image characteristic vector set in each hyperspectral original image by taking the K moment as a main component; a first function module for calculating a bilateral empirical cumulative distribution function based on the set of image feature vectors; the second function module is used for calculating an experience Copula function based on the bilateral experience cumulative distribution function; the probability estimation module is used for estimating bilateral tail end probability values which are distributed in a combined mode in all dimensions through an empirical Copula function; the result analysis module is used for obtaining the analysis result of overexposure, overdose or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides; and the automatic adjustment module is used for adjusting the exposure time according to the abnormal exposure condition until the exposure is normal.
Working principle: compared with the traditional automatic exposure of the visible light image, the automatic exposure of the hyperspectral image involves the characteristics of numerous image spectrum, large data volume, obvious imaging difference of different spectrum segments and the like; according to the application, the automatic exposure function of the hyperspectral camera is efficiently completed by utilizing the feature vectors corresponding to the original images under different exposure time, and good accuracy and generalization can be ensured by utilizing the feature vectors of the original images and the experience coupler function; in addition, in the application process, the application can automatically adjust the exposure time by combining the returned result without adding additional detection equipment and detection equipment, realizes the automatic exposure function, and has strong adaptability and practicability to application scenes.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the application, and is not meant to limit the scope of the application, but to limit the application to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the application are intended to be included within the scope of the application.
Claims (6)
1. The hyperspectral image automatic exposure method based on the COPOD algorithm is characterized by comprising the following steps of:
acquiring a hyperspectral original image with exposure time;
calculating an image feature vector set in each hyperspectral original image by taking the K moment as a main component;
calculating a bilateral experience cumulative distribution function based on the image feature vector set;
calculating an empirical Copula function based on the bilateral empirical cumulative distribution function;
estimating bilateral tail probability values of joint distribution in all dimensions through an empirical Copula function;
obtaining an analysis result of overexposure, darkness or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides;
adjusting exposure time for abnormal exposure condition until exposure is normal;
the calculation process of the image feature vector set specifically comprises the following steps:
calculating an index set corresponding to the pixel value statistic from band to band by taking the K moment as a main component;
the pixel value statistics include: a first-order origin moment, a second-order center moment, a third-order center distance, a fourth-order center moment, an entropy value and a pixel proportion with brightness larger than a given value of each channel pixel value;
the calculation process of the bilateral experience cumulative distribution function specifically comprises the following steps:
the calculation formula for calculating the left cumulative probability density function is specifically:
the right cumulative probability density function is calculated, and the calculation formula is specifically as follows:
calculating a skewness coefficient, wherein a calculation formula specifically comprises:
wherein ,a left cumulative probability density function representing a d-th dimensional feature; />A right cumulative probability density function representing a d-th dimensional feature; b d A skewness factor representing a d-th dimensional feature; x represents the order statistic of the samples; x is X i Representing the actual value of the ith sample; />Representing a sample mean; n represents the selected sample size;
the calculation process of the bilateral tail end probability value specifically comprises the following steps:
from the hyperspectral image samples, the left tail probability is calculated:
from the hyperspectral image samples, the right tail probability is calculated:
calculating the integral tail probability according to the hyperspectral image sample:
wherein ,pl Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability; d represents a dimension feature sequence number;left experience representing jth dimension featureA value; />Right experience values representing the j-th dimensional feature; representation->Observation statistics of the j-th dimension feature;
the analysis result is obtained by the following steps:
obtaining a judgment index score according to analysis of the probability values of the two tail ends, wherein a calculation formula of the index score is specifically as follows:
S(x i )＝max(p l ，p r ，p s )
wherein ,S(xi ) Representing a decision index score; p is p l Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability;
and repeatedly calculating to obtain the judgment index scores of all samples, and analyzing and judging overexposure, overdrising or normal exposure of the hyperspectral original image according to the distribution interval of the judgment index scores.
2. The automatic exposure method for hyperspectral image based on COPOD algorithm as claimed in claim 1, wherein the acquisition process of hyperspectral original image is specifically as follows:
collecting an original hyperspectral image with exposure time in a line scanning type or snapshot type by adopting a hyperspectral camera;
clipping the original hyperspectral image into an image with uniform height, width and spectrum channel number;
and naming the cut images in a unified format and recording corresponding exposure time to obtain hyperspectral original images.
3. The hyperspectral image automatic exposure method based on the COPOD algorithm as claimed in claim 1, wherein the calculation formula of the entropy value is specifically as follows:
wherein X represents a random variable of pixel gray values of a certain frame of the hyperspectral image; entopy (X) represents the entropy value of the random variable X; x is X i Representing the actual value of the ith sample; p (x=x) i ) Indicating that the sample has a value of X i Is a frequency of (2); n represents the total number of samples; n represents the selected sample size.
4. The method for automatically exposing hyperspectral images based on the COPOD algorithm according to claim 1, wherein the pixel value statistics further comprise combined statistics consisting of at least two of a first order origin moment, a second order center moment, a third order center distance, a fourth order center moment, an entropy value, and a pixel proportion with brightness greater than a given value.
5. The hyperspectral image automatic exposure method based on the COPOD algorithm as claimed in claim 1, wherein the calculation process of the empirical Copula function is specifically as follows:
left-hand empirical Copula observations were calculated from hyperspectral image samples:
the right-hand empirical Copula observation was calculated from the hyperspectral image samples:
calculating Copula observation statistics: if b d < 0, thenOtherwise, go (L)>
wherein ,represent left experience values; />Represent the right experience value; />Representing Copula observations; x is x i A sequence statistic representing a number i; p is p d A skewness factor representing a d-th dimensional feature; />Representing observed statistics.
6. The hyperspectral image automatic exposure system based on the COPOD algorithm is characterized by comprising:
the image acquisition module is used for acquiring a hyperspectral original image with exposure time;
the characteristic calculation module is used for calculating an image characteristic vector set in each hyperspectral original image by taking the K moment as a main component;
a first function module for calculating a bilateral empirical cumulative distribution function based on the set of image feature vectors;
the second function module is used for calculating an experience Copula function based on the bilateral experience cumulative distribution function;
the probability estimation module is used for estimating bilateral tail end probability values which are distributed in a combined mode in all dimensions through an empirical Copula function;
the result analysis module is used for obtaining the analysis result of overexposure, overdose or normal exposure of the hyperspectral original image according to the probability analysis of the tail ends of the two sides;
the automatic adjusting module is used for adjusting the exposure time according to the abnormal exposure condition until the exposure is normal;
the calculation process of the image feature vector set specifically comprises the following steps:
calculating an index set corresponding to the pixel value statistic from band to band by taking the K moment as a main component;
the pixel value statistics include: a first-order origin moment, a second-order center moment, a third-order center distance, a fourth-order center moment, an entropy value and a pixel proportion with brightness larger than a given value of each channel pixel value;
the calculation process of the bilateral experience cumulative distribution function specifically comprises the following steps:
the calculation formula for calculating the left cumulative probability density function is specifically:
the right cumulative probability density function is calculated, and the calculation formula is specifically as follows:
calculating a skewness coefficient, wherein a calculation formula specifically comprises:
wherein ,a left cumulative probability density function representing a d-th dimensional feature; />A right cumulative probability density function representing a d-th dimensional feature; b d A skewness factor representing a d-th dimensional feature; x represents the order statistic of the samples; x is X i Representing the actual value of the ith sample; />Representing a sample mean; n represents the selected sample size;
the calculation process of the bilateral tail end probability value specifically comprises the following steps:
according to hyperspectralImage sample, calculating left tail probability:
from the hyperspectral image samples, the right tail probability is calculated:
calculating the integral tail probability according to the hyperspectral image sample:
wherein ,pl Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability; d represents a dimension feature sequence number;left experience values representing the j-th dimensional feature; />Right experience values representing the j-th dimensional feature; representation->Observation statistics of the j-th dimension feature;
the analysis result is obtained by the following steps:
obtaining a judgment index score according to analysis of the probability values of the two tail ends, wherein a calculation formula of the index score is specifically as follows:
S(x i )＝max(p l ，p r ，p s )
wherein ,S(xi ) Representing a decision index score; p is p l Represents the left tail probability; p is p r Represents the right tail probability; p is p s Representing the overall tail probability;
and repeatedly calculating to obtain the judgment index scores of all samples, and analyzing and judging overexposure, overdrising or normal exposure of the hyperspectral original image according to the distribution interval of the judgment index scores.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210415974.0A CN114785965B (en) | 2022-04-20 | 2022-04-20 | Automatic hyperspectral image exposure method and system based on COPOD algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210415974.0A CN114785965B (en) | 2022-04-20 | 2022-04-20 | Automatic hyperspectral image exposure method and system based on COPOD algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114785965A CN114785965A (en) | 2022-07-22 |
| CN114785965B true CN114785965B (en) | 2023-09-05 |
Family
ID=82431200
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210415974.0A Active CN114785965B (en) | 2022-04-20 | 2022-04-20 | Automatic hyperspectral image exposure method and system based on COPOD algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114785965B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106791470A (en) * | 2016-12-28 | 2017-05-31 | 上海兴芯微电子科技有限公司 | Exposal control method and device based on HDR camera head |
| WO2018085841A1 (en) * | 2016-11-07 | 2018-05-11 | BioSensing Systems, LLC | Calibration method and apparatus for active pixel hyperspectral sensors and cameras |
| CN110570387A (en) * | 2019-09-16 | 2019-12-13 | 江南大学 | image fusion method based on feature level Copula model similarity |
| CN112989338A (en) * | 2021-01-04 | 2021-06-18 | 腾讯科技（深圳）有限公司 | Abnormal application data detection method and device, electronic equipment and storage medium |
| CN113761292A (en) * | 2021-04-29 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Object identification method and device, computer equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FI113897B (en) * | 2001-11-23 | 2004-06-30 | Planmed Oy | Automatic exposure procedure and automatic exposure system |
| CA2773925A1 (en) * | 2012-04-10 | 2013-10-10 | Robert Kendall Mcconnell | Method and systems for computer-based selection of identifying input for class differentiation |
-
2022
- 2022-04-20 CN CN202210415974.0A patent/CN114785965B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018085841A1 (en) * | 2016-11-07 | 2018-05-11 | BioSensing Systems, LLC | Calibration method and apparatus for active pixel hyperspectral sensors and cameras |
| CN106791470A (en) * | 2016-12-28 | 2017-05-31 | 上海兴芯微电子科技有限公司 | Exposal control method and device based on HDR camera head |
| CN110570387A (en) * | 2019-09-16 | 2019-12-13 | 江南大学 | image fusion method based on feature level Copula model similarity |
| CN112989338A (en) * | 2021-01-04 | 2021-06-18 | 腾讯科技（深圳）有限公司 | Abnormal application data detection method and device, electronic equipment and storage medium |
| CN113761292A (en) * | 2021-04-29 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Object identification method and device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 孙旭东 ; 刘燕德 ; 肖怀春 ; 张智诚 ; 李泽敏 ; 吕强 ; .正常、缺素和黄龙病柑桔叶片高光谱成像快速诊断.光谱学与光谱分析.2017,(第02期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114785965A (en) | 2022-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | An imaging-inspired no-reference underwater color image quality assessment metric | |
| CN109035166B (en) | Electrical equipment infrared image enhancement method based on non-subsampled shear wave transformation | |
| CN112379231B (en) | Equipment detection method and device based on multispectral image | |
| CN107798652A (en) | Image processing method, device, readable storage medium storing program for executing and electronic equipment | |
| CN112381784A (en) | Equipment detecting system based on multispectral image | |
| DE112016002322T5 (en) | Image processing apparatus, image processing method, program, program recording recording medium, image capturing apparatus, and image recording / reproducing apparatus | |
| CN107993209A (en) | Image processing method, device, computer-readable recording medium and electronic equipment | |
| Okarma | Current trends and advances in image quality assessment | |
| RU2419150C1 (en) | Method to process sequence of images to detect and follow air objects | |
| CN114785965B (en) | Automatic hyperspectral image exposure method and system based on COPOD algorithm | |
| CN112507818A (en) | Illumination estimation method and system based on near-infrared image | |
| CN110223253B (en) | Defogging method based on image enhancement | |
| CN113252207B (en) | Electrical equipment metal surface temperature difference detection method and system | |
| TWM458747U (en) | Image processing module | |
| Yang et al. | Blind image quality assessment on authentically distorted images with perceptual features | |
| CN110020999B (en) | Uncooled infrared thermal image self-adaptive mapping method based on homomorphic filtering | |
| US8866972B1 (en) | Method for transmitting spectrum information | |
| Thakur et al. | Classification of color hazy images | |
| CN114140742A (en) | Track foreign matter intrusion detection method based on light field depth image | |
| CN113436096A (en) | Push-broom hyperspectral imaging strip noise elimination method based on pixel calibration | |
| CN111986109A (en) | Remote sensing image defogging method based on full convolution network | |
| CN114240776B (en) | Demosaicing and compression fusion framework for MSFA hyperspectral image | |
| Kareem et al. | No reference quality of the hazy images depending on transmission component estimation | |
| CN117474815B (en) | Hyperspectral image calibration method and system | |
| CN110706165B (en) | Underwater image dodging algorithm based on EME and Mask |
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 |