A multi-spectral radiation temperature measurement method based on four sCMOS cameras

Abstract

The present invention relates to the technical field of optical elements, systems or instruments, and specifically to a multi-spectral radiation temperature measurement method based on four sCMOS cameras, which solves the technical problems of short response time or low measurement accuracy of existing temperature measurement methods. The method of the present invention constructs a multi-spectral radiation temperature measurement system, which includes a temperature field, four sets of filters, four back-illuminated sCMOS cameras, a black body furnace and a host computer. The camera is calibrated by the black body furnace, and the four cameras collect two-dimensional spectral images of the temperature field. The host computer processes the collected two-dimensional spectral images; the temperature field is obtained by inversion using the Powell constrained optimization algorithm. The method of the present invention finally uses the Powell optimization algorithm to convert the temperature solution problem into a minimum value solution problem, which can eliminate the influence of the emissivity, and combines the advantages of multi-spectral point temperature measurement and array temperature measurement, so as to achieve high-precision solution of the two-dimensional temperature field.

Description

Technical Field

[0001] The present invention relates to the technical field of optical elements, systems or instruments, and in particular to multi-spectral radiation temperature measurement, and specifically to a multi-spectral radiation temperature measurement method based on four sCMOS cameras. Background Art

[0002] High-precision measurement of temperature distribution plays an important guiding role in the development of aircraft engines and the assessment of weapon damage effects. For example, the temperature of the combustion chamber of an aircraft engine will directly affect the thrust of the rocket; the deflagration temperature refers to the highest temperature that an explosive can produce during an explosion, which can indirectly indicate the thermal damage effect of the explosive.

[0003] At present, the common temperature measurement methods are mainly divided into contact and non-contact. The contact temperature measurement method is simple to operate, with high test accuracy and reliability, but it requires the sensor to invade the temperature field to be measured, which will affect the flow field. Conventional materials are difficult to survive during high-temperature measurements, and contact temperature sensors usually have a slow response time, making it difficult to test transient temperature fields. Non-contact temperature measurement will not interfere with the explosion flow field. With the development of devices, the response speed has been greatly improved, but the temperature measurement accuracy of existing infrared radiation temperature measurement, single-wavelength temperature measurement, and multi-wavelength temperature measurement methods is greatly affected by the emissivity of the object being measured. Summary of the invention

[0004] In order to overcome the technical defects of the existing temperature measurement methods, such as short response time or low measurement accuracy, the present invention provides a multi-spectral radiation temperature measurement method based on four sCMOS cameras.

[0005] The present invention provides a multi-spectral radiation temperature measurement method based on four sCMOS cameras, comprising the following steps:

[0006] S1. According to the temperature range of the measured temperature field, four sets of filters and four back-illuminated sCMOS cameras are selected. The four sets of filters are used to obtain spectral information of specific bands.

[0007] S2. Calibrate four back-illuminated sCMOS cameras using a blackbody furnace to establish a functional relationship between camera grayscale and temperature.

[0008] According to Planck's blackbody radiation law, when the radiation emitted by an ideal blackbody point light source is uniformly distributed on a spherical surface, the radiation intensity is the power passing through a unit area, and the radiant brightness is defined as the radiant power per unit area and unit solid angle in a specific direction; according to Lambert's cosine law, when uniformly emitted from a surface element, the radiant brightness in any direction is The radiation intensity perpendicular to the surface element The relationship between As shown:

[0009] ,

[0010] Therefore, the blackbody radiance in Planck's blackbody radiation law is Expressed as a formula :

[0011] ;

[0012] In the wavelength range where the back-illuminated sCMOS camera works normally, the received radiance is linearly related to the output signal intensity. The linear relationship is as follows: As shown:

[0013] ,

[0014] formula middle, Back-illuminated sCMOS camera at wavelength The signal strength of the output is Indicates wavelength The blackbody radiance corresponding to the temperature T, constant A and constant B are obtained through calibration experiments; wavelength It is within the wavelength range where back-illuminated sCMOS cameras can work properly.

[0015] First, set the blackbody temperature to , the wavelength can be obtained through the blackbody furnace calibration experiment Lower temperature The corresponding blackbody radiance The gray value of the output The relationship is the formula :

[0016] ;

[0017] S3, four back-illuminated sCMOS cameras and four sets of filters are respectively used to collect two-dimensional spectral images of the temperature field;

[0018] S4, the host computer processes the collected two-dimensional spectral image of the temperature field; the temperature field is inverted using the Powell constrained optimization algorithm;

[0019] According to Planck's blackbody radiation law and emissivity theory, assuming that The i-th wavelength channel λ at temperature i The output signal is grayscale g i , channel λ iThe corresponding spectral emissivity is , then the non-blackbody radiance is expressed by the formula calculate,

[0020] ;

[0021] According to the formula and formula The formula can be obtained :

[0022] ;

[0023] Since the spectral emissivity of a black body is 1, then the grayscale of the object being measured is as follows As shown:

[0024] ;

[0025] formula According to the grayscale output at each wavelength , the emissivity is corrected by a constrained optimization algorithm , the real temperature of the object under test and the emissivity distribution at different wavelengths are obtained by inversion;

[0026] According to the formula , when the emissivity is determined , then the temperature of each channel should be consistent and equal to the true temperature, so the theoretical deviation of the temperature calculated by each channel should tend to 0, as shown in the formula As shown:

[0027] ∑ i = 1 n [ T i − T ¯ ] 2 = 0 ;

[0028] formula middle, Calculate the average temperature for all channels and transform the solution of the true temperature of each channel into a minimum optimization problem, as shown in the formula As shown:

[0029] m i n f ( x ) = ∑ i = 1 n [ T i − T ¯ ] 2 → 0 ;

[0030] The actual temperature of each channel can be calculated by solving the minimum optimization problem using the Powell optimization algorithm.

[0031] Preferably, in the Powell optimization algorithm, is the multivariable objective function to be optimized, where , since the flame emission rate is less than 1, add constraints , then the sub-steps of the Powell optimization algorithm are:

[0032] S41. Initial direction set: Select the initial point and a set of linearly independent nonzero vectors , select the standard orthogonal basis as the initial vector and set the error limit to , used to determine when the Powell optimization algorithm stops and to set the iteration counter ;

[0033] S42, Directional search: follow the In each direction Perform a one-dimensional search to find a scalar in each direction , so that Minimum; Update for , repeat the current process for all directions in turn;

[0034] S43, update search direction: After searching all directions, use Replace the direction that minimizes the function value, thereby obtaining a new set of search directions; is the point before the start of this iteration, is the point after all directions have been searched;

[0035] S44, iteration and convergence: repeat the above step S43 and update the iteration point is obtained after the current iteration , until the objective function value is less than the threshold, and the change value of the objective function between two iterations is less than the preset threshold.

[0036] Preferably, in step S42, check each updated ,when When the constraint boundary is exceeded, it is set to the boundary value. At the next update, if you want to continue updating , then perform the "overflow" operation to update it from the boundary value on the other side.

[0037] Preferably, the wavelengths of the four groups of filters are 510 nm, 521 nm, 694 nm and 768 nm respectively, and the blackbody furnace is used for calibration experiments at 600-1000°C.

[0038] Preferably, in step S4, first, the host computer needs to perform registration processing on the collected two-dimensional spectral image of the temperature field, and the registration processing includes: denoising the image to be registered, then extracting grayscale gradient features from the image to be registered, using the extracted grayscale gradient features, and using global matching to calculate the mutual correlation coefficient between the two sets of features for feature matching; estimating affine transformation based on feature points, and using affine transformation to transform the image to be registered into the reference image coordinate system; using the nearest interpolation method to calculate the pixel values ​​of non-integer coordinate positions to complete the registration.

[0039] Compared with the prior art, the technical solution provided by the present invention has the following technical effects: in the method described in the present invention, four sCMOS cameras are combined with four groups of filters, a blackbody furnace, and a host computer to construct a multi-spectral radiation temperature measurement system. The optical system is simple and the cost is low. The temperature information and spectral information of the temperature field can be obtained simultaneously. The method described in the present invention finally uses the Powell optimization algorithm to convert the temperature solution problem into a minimum value solution problem, which can eliminate the influence of the emissivity, combines the advantages of multi-spectral point temperature measurement and array temperature measurement, and can achieve high-precision solution of the two-dimensional temperature field. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 Schematic diagram of the structure of a multi-spectral radiation temperature measurement system based on four sCMOS cameras described in an embodiment of the present invention;

[0043] Figure 2 This is a flow chart of a multi-spectral radiation temperature measurement method based on four sCMOS cameras described in an embodiment of the present invention;

[0044] Figure 3 This is a relationship diagram between the radiance and grayscale of a filter with a wavelength of 510 nm in an embodiment of the present invention;

[0045] Figure 4 This is a relationship diagram between the radiance and grayscale of a filter with a wavelength of 521 nm in an embodiment of the present invention;

[0046] Figure 5 This is a relationship diagram between the radiance and grayscale of a filter with a wavelength of 694 nm in an embodiment of the present invention;

[0047] Figure 6 This is a relationship diagram between the radiance and grayscale of a filter with a wavelength of 768nm in an embodiment of the present invention;

[0048] Figure 7 It is a schematic diagram of the measurement results of the two-dimensional temperature field distribution of the flame in a certain embodiment of the present invention.

[0049] In the figure: 1. Temperature field; 2. Filter; 3. Back-illuminated sCMOS camera; 4. Blackbody furnace; 5. Host computer. DETAILED DESCRIPTION

[0050] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the scheme of the present invention will be further described below. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present invention, rather than all of the embodiments.

[0052] The following is combined with Figures 1 to 7 The specific embodiments of the present invention are described in detail.

[0053] In one embodiment, Figure 1 As shown, the names include:

[0054] Based on the above embodiments, in a preferred embodiment, a multi-spectral radiation temperature measurement method based on four sCMOS cameras includes the following steps:

[0055] S1. According to the temperature range of the measured temperature field 1, four sets of filters 2 and four back-illuminated sCMOS cameras 3 are selected. The four sets of filters 2 are used to obtain spectral information of specific bands. The wavelengths of the four sets of filters 2 are 510 nm, 521 nm, 694 nm and 768 nm respectively, in order to obtain spectral information of specific bands.

[0056] S2. Calibrate four back-illuminated sCMOS cameras 3 at 600-1000° C. using a black body furnace 4 to establish a functional relationship between camera grayscale and temperature. In a specific embodiment, the quantum efficiency of the four back-illuminated sCMOS cameras 3 in the 500-800 nm band is required to be no less than 40%.

[0057] According to Planck's blackbody radiation law, when the radiation emitted by an ideal blackbody point light source is uniformly distributed on a spherical surface, the radiation intensity is the power passing through a unit area, and the radiant brightness is defined as the radiant power per unit area and unit solid angle in a specific direction; according to Lambert's cosine law, when uniformly emitted from a surface element, the radiant brightness in any direction is The radiation intensity perpendicular to the surface element The relationship between As shown:

[0058] ,

[0059] Therefore, the blackbody radiance in Planck's blackbody radiation law is Expressed as a formula :

[0060] ;

[0061] According to the relationship between emissivity and wavelength, radiators are divided into three types: black body, gray body and selective radiator. Thermal radiation theory takes black body with emissivity of 1 as the research object. Gray body regards the measured object as one with emissivity less than 1, but does not change with wavelength. In actual tests, the measured objects are all radiators with emissivity changing with wavelength.

[0062] In order to solve the problem that the temperature and emissivity of the flame are difficult to measure, a multi-spectral radiation temperature measurement algorithm based on average temperature is used to reconstruct the emissivity distribution of the flame at different wavelengths while measuring the flame temperature distribution; within the wavelength range of normal operation of the back-illuminated sCMOS camera 3, the received radiance and its output signal intensity are linearly related, and the linear relationship is as shown in the formula As shown:

[0063] ,

[0064] formula middle, Back-illuminated sCMOS camera 3 at wavelength The signal strength of the output is Indicates wavelength The radiance corresponding to the temperature T, constant A and constant B are obtained through calibration experiments; wavelength The wavelength range of the back-illuminated sCMOS camera 3 is within normal working range. Otherwise, due to the influence of the sensor quantum efficiency, and It shows nonlinear relationship;

[0065] First, set the blackbody temperature to , the wavelength can be obtained through the blackbody furnace 4 calibration experiment Lower temperature The corresponding blackbody radiance The gray value of the output The relationship is the formula :

[0066] ;

[0067] S3, four back-illuminated sCMOS cameras 3 and four sets of filters 2 are respectively combined to collect two-dimensional spectral images of the temperature field 1;

[0068] S4, the host computer 5 processes the collected two-dimensional spectrum image of the temperature field 1; and inverts the temperature field 1 using the Powell constrained optimization algorithm;

[0069] First, the host computer 5 is required to perform registration processing on the collected two-dimensional spectral image of the temperature field 1, and the registration processing includes: denoising the image to be registered, extracting grayscale gradient features from the image to be registered, using the extracted grayscale gradient features, using global matching to calculate the mutual correlation coefficient between two sets of grayscale gradient features for feature matching; performing affine transformation estimation based on feature points, and using affine transformation to transform the image to be registered into the reference image coordinate system; using the neighbor interpolation method to calculate the pixel value of the non-integer coordinate position to complete the registration;

[0070] Secondly, according to Planck's blackbody radiation law and emissivity theory, assuming that The i-th wavelength channel λ at temperature i The output signal is grayscale g i , channel λ i The corresponding spectral emissivity is , then the non-blackbody radiance is expressed by the formula calculate,

[0071] ;

[0072] According to the formula and formula The formula can be obtained :

[0073] ;

[0074] Since the spectral emissivity of a black body is 1, then the grayscale of the object being measured is as follows As shown:

[0075] ;

[0076] formula According to the grayscale output at each wavelength , the emissivity is corrected by a constrained optimization algorithm , the real temperature of the object under test and the emissivity distribution at different wavelengths are obtained by inversion;

[0077] According to the formula , when the emissivity is determined , then the temperature of each channel should be consistent and equal to the true temperature, so the theoretical deviation of the temperature calculated by each channel should tend to 0, as shown in the formula As shown:

[0078] ∑ i = 1 n [ T i − T ¯ ] 2 = 0 ;

[0079] formula middle, Calculate the average temperature for all channels and transform the solution of the true temperature of each channel into a minimum optimization problem, as shown in the formula As shown:

[0080] m i n f ( x ) = ∑ i = 1 n [ T i − T ¯ ] 2 → 0 ;

[0081] The Powell optimization algorithm is used to solve the minimum optimization problem, and the true temperature of each channel can be calculated;

[0082] In the Powell optimization algorithm, let is the multivariable objective function to be optimized, where Since the flame emissivity is less than 1, in order to improve the temperature inversion accuracy and speed of the Powell optimization algorithm, the constraint condition is added , then the sub-steps of the Powell optimization algorithm are:

[0083] S41. Initial direction set: Select the initial point and a set of linearly independent nonzero vectors , select the standard orthogonal basis as the initial vector and set the error limit to , used to determine when the Powell optimization algorithm stops and to set the iteration counter ;

[0084] S42, Directional search: follow the In each direction Perform a one-dimensional search to find a scalar in each direction , so that Minimum; Update for , repeat the current process for all directions in turn; check after each update ,when When the constraint boundary is exceeded, it is set to the boundary value. At the next update, if you want to continue updating , then perform the "overflow" operation to update from the boundary value on the other side;

[0085] S43, update search direction: After searching all directions, use Replace the direction that minimizes the function value, thereby obtaining a new set of search directions; is the point before the start of this iteration, is the point after all directions have been searched;

[0086] S44, iteration and convergence: repeat the above step S43 and update the iteration point is obtained after the current iteration , until the objective function value is less than the threshold, and the change value of the objective function between two iterations is less than the preset threshold.

[0087] The above is only a specific implementation of the present invention, which enables those skilled in the art to understand or implement the present invention. Although detailed descriptions are given with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments, and they should all be covered by the protection scope of the claims.

Claims

1. A multi-spectral radiation temperature measurement method based on four sCMOS cameras, characterized in that: The steps include: S1. Select four sets of filters (2) and four back-illuminated sCMOS cameras (3) according to the temperature range of the measured temperature field (1), wherein the four sets of filters (2) are used to obtain spectral information of a specific band; S2, conducting a calibration experiment on four back-illuminated sCMOS cameras (3) using a black body furnace (4) to establish a functional relationship between camera grayscale and temperature; According to Planck's blackbody radiation law, when the radiation emitted by an ideal blackbody point light source is uniformly distributed on a spherical surface, the radiation intensity is the power passing through a unit area, and the radiant brightness is defined as the radiant power per unit area and unit solid angle in a specific direction; according to Lambert's cosine law, when uniformly emitted from a surface element, the radiant brightness in any direction is The radiation intensity perpendicular to the surface element The relationship between As shown: ， Therefore, the blackbody radiance in Planck's blackbody radiation law is Expressed as a formula : ； Within the wavelength range of normal operation of the back-illuminated sCMOS camera (3), the received radiance is linearly related to the output signal intensity. The linear relationship is as follows: As shown: , formula middle, Back-illuminated sCMOS camera (3) at wavelength The signal strength of the output is Indicates wavelength The blackbody radiance corresponding to the temperature T, constant A and constant B are obtained through calibration experiments; wavelength Within the wavelength range where the back-illuminated sCMOS camera (3) can operate normally; First, set the blackbody temperature to , the wavelength is obtained by the blackbody furnace (4) calibration experiment Lower temperature The corresponding blackbody radiance The gray value of the output The relationship is the formula : ； S3, four back-illuminated sCMOS cameras (3) and four sets of filters (2) are respectively combined to collect two-dimensional spectral images of the temperature field (1); S4, the host computer (5) processes the collected two-dimensional spectral image of the temperature field (1); and inverts the temperature field (1) using the Powell constrained optimization algorithm; According to Planck's blackbody radiation law and emissivity theory, assuming that The i-th wavelength channel λ at temperature i The output signal is grayscale g i , channel λ i The corresponding spectral emissivity is , then the non-blackbody radiance is expressed by the formula calculate, ； According to the formula and formula The formula can be obtained : ； Since the spectral emissivity of a black body is 1, then the grayscale of the object being measured is as follows As shown: ； formula According to the grayscale output at each wavelength , the emissivity is corrected by a constrained optimization algorithm , the real temperature of the object under test and the emissivity distribution at different wavelengths are obtained by inversion; According to the formula , when the emissivity is determined , then the temperature of each channel should be consistent and equal to the true temperature, so the theoretical deviation of the temperature calculated by each channel should tend to 0, as shown in the formula As shown: ； formula middle, Calculate the average temperature for all channels and transform the solution of the true temperature of each channel into a minimum optimization problem, as shown in the formula As shown: ； The actual temperature of each channel can be calculated by solving the minimum optimization problem using the Powell optimization algorithm.

2. The multi-spectral radiation temperature measurement method based on four sCMOS cameras according to claim 1, characterized in that: In the Powell optimization algorithm, let is the multivariable objective function to be optimized, where , since the flame emission rate is less than 1, add constraints , then the sub-steps of the Powell optimization algorithm are: S41. Initial direction set: Select the initial point and a set of linearly independent nonzero vectors , select the standard orthogonal basis as the initial vector and set the error limit to , used to determine when the Powell optimization algorithm stops and to set the iteration counter ; S42, Directional search: follow the In each direction Perform a one-dimensional search to find a scalar in each direction , so that Minimum; Update for , repeat the current process for all directions in turn; S43, update search direction: After searching all directions, use Replace the direction that minimizes the function value, thereby obtaining a new set of search directions; is the point before the start of this iteration, is the point after all directions have been searched; S44, iteration and convergence: repeat the above step S43 and update the iteration point is obtained after the current iteration , until the objective function value is less than the threshold, and the change value of the objective function between two iterations is less than the preset threshold.

3. The multi-spectral radiation temperature measurement method based on four sCMOS cameras according to claim 2 is characterized in that: In step S42, check each updated ,when When the constraint boundary is exceeded, it is set to the boundary value. At the next update, if you want to continue updating , then perform the "overflow" operation to update it from the boundary value on the other side.

4. A multi-spectral radiation temperature measurement method based on four sCMOS cameras according to any one of claims 1 to 3, characterized in that: The wavelengths of the four sets of filters (2) are 510 nm, 521 nm, 694 nm and 768 nm respectively, and the black body furnace (4) is subjected to a calibration experiment at 600-1000° C.

5. The multi-spectral radiation temperature measurement method based on four sCMOS cameras according to claim 4 is characterized in that: In step S4, first, the host computer (5) is required to perform registration processing on the collected two-dimensional spectral image of the temperature field (1), and the registration processing includes: performing denoising processing on the image to be registered, then extracting grayscale gradient features from the image to be registered, using the extracted grayscale gradient features, using global matching to calculate the mutual correlation coefficient between two groups of grayscale gradient features for feature matching; performing affine transformation estimation based on feature points, using affine transformation to transform the image to be registered into the reference image coordinate system; using the nearest interpolation method to calculate the pixel value of the non-integer coordinate position to complete the registration.

Images