Infrared image temperature calibration method and system

Abstract

The invention relates to an infrared image temperature calibration method and system. Error amount generated by distance measurement between an infrared radiation source and an infrared imager is determined through atmospheric transmittance. And calling object surface emissivity preliminary information to reduce the surface emissivity error amount of the infrared radiation source in the actual measurement process. Based on the atmospheric transmittance, a temperature and hyper-parameter joint estimation model can be formed by establishing an infrared radiation model and constructing a forward model. A Bayesian inference method is utilized to deduce the surface emissivity of an object, meanwhile, prior information of different measurement materials is considered, and model parameters and temperature distribution are reconstructed by combining an improved model with a maximum posteriori estimation method. The improved infrared thermal radiation model is suitable for scenes with different measurement distances, humidity factors and emissivity. Through verification of multiple experiments of the black body and the high-speed direct-driven fan, the method can significantly reduce the measurement error and improve the accuracy of temperature measurement.

Description

Technical Field

[0001] The present application relates to the field of infrared image technology, and in particular to a method and system for infrared image temperature calibration. Background Art

[0002] Infrared temperature measurement technology can quickly obtain the surface temperature of an object through non-contact measurement. Therefore, infrared temperature measurement technology is widely used in medical, industrial, transportation, construction, security and other fields. However, in practical applications, the accuracy of infrared temperature measurement is often affected by many factors such as ambient temperature, humidity, atmospheric attenuation, emissivity of the object material, measurement distance and sensor performance, resulting in the measured temperature often being lower than the actual temperature. Therefore, it is necessary to propose a method and system for infrared image temperature calibration to address the low accuracy of traditional infrared image temperature measurement. Summary of the invention

[0003] Based on this, it is necessary to propose a method and system for infrared image temperature calibration to address the defect of low accuracy of traditional infrared image temperature measurement.

[0004] The present application provides a method for infrared image temperature calibration, comprising:

[0005] Determine atmospheric transmittance;

[0006] Based on the atmospheric transmittance, an infrared radiation model is established;

[0007] Call the object surface to emit the first-come-first-served information;

[0008] Build a forward model;

[0009] Based on the infrared radiation model, the a priori information of the object surface emission and the forward model, a joint estimation model of temperature and hyperparameters is formed;

[0010] An alternating iterative optimization algorithm is used in the joint estimation model of temperature and hyperparameters to obtain the final temperature result.

[0011] Furthermore, the relative humidity, measurement distance and absorption characteristic data of the infrared band are obtained;

[0012] Incorporate relative humidity, measurement distance, and infrared absorption characteristic data into the parameter formula of the Beer-Lambert law;

[0013] Based on the parameter formula of the Beer-Lambert law, calculate the attenuation data of infrared radiation in the atmosphere;

[0014] Using the attenuation data of infrared radiation in the atmosphere, the expression of atmospheric transmittance is determined.

[0015] Furthermore, the expression of atmospheric transmittance is:

[0016]

[0017] Where τ is the atmospheric transmittance, k(h w ) is the water vapor content h w The relevant attenuation coefficient, d is the propagation distance of infrared radiation, and e is a natural constant.

[0018] Furthermore, the relationship between the attenuation coefficient and temperature and humidity is:

[0019]

[0020] in, is the temperature T 1 The amount of water vapor condensed, is the temperature T 2 The amount of water vapor condensed, and is the relative humidity at different temperatures, At temperature T 1 The attenuation coefficient when At temperature T 2 The attenuation coefficient when .

[0021] Furthermore, emissivity is defined as a parameter to be attached with a value;

[0022] Call the initial infrared radiation model;

[0023] The emissivity is incorporated into the initial infrared radiation model to form an infrared radiation model.

[0024] Furthermore, the expression of the temperature measured by the infrared imager in the infrared radiation model is:

[0025]

[0026] Among them, T m is the temperature measured by the infrared imager, T r is the actual temperature of the target object, ε is the surface emissivity of the object, T u is the background temperature, T a is the atmospheric temperature, and n is a constant related to the infrared sensor material.

[0027] Furthermore, the surface emissivity of the object is defined as a parameter to be attached with a value;

[0028] Receive experimental measurement data of the surface emissivity of the object;

[0029] Statistics experimental measurement data;

[0030] Using Gaussian distribution, the mean and variance of the surface emissivity of the object in the experimental measurement data are determined.

[0031] Furthermore, the forward model of the temperature measured by the infrared imager is expressed as:

[0032] T m =H(T r ,ε)+ζ

[0033] Among them, ζ is the measured swing amount;

[0034] Based on the inverse gamma distribution, the variance of the measured swing is predetermined;

[0035] The variance of the actual temperature distribution of the target object is defined as an inverse gamma distribution.

[0036] Further, based on the actual temperature of the target object, the surface emissivity of the object, a first hidden variable related to the actual temperature of the target object, and a second hidden variable related to the surface emissivity of the object, a joint estimation model of temperature and hyperparameters is established;

[0037] Use JMAP estimation method to optimize the joint estimation model of temperature and hyperparameters;

[0038] Generates a cost function for the joint estimation model of temperature and hyperparameters.

[0039] The present application provides a system for infrared image temperature calibration, comprising:

[0040] A processing end, used to execute the infrared image temperature calibration method;

[0041] An external device is communicatively connected to the processing end, and the external device includes one or more of an infrared camera and a parameter collector.

[0042] The present application relates to a method and system for infrared image temperature calibration, which determines the error amount caused by the measured distance between the infrared radiation source and the infrared imager through atmospheric transmittance. The surface emissivity error of the infrared radiation source is reduced by calling the prior information of the object surface emission during the actual measurement process. Based on the atmospheric transmittance, an infrared radiation model is established and a forward model is constructed to form a joint estimation model of temperature and hyperparameters. The surface emissivity of the object is derived using the Bayesian inference method, and the prior information of different measurement materials is considered at the same time. Combined with the improved model, the model parameters and temperature distribution are reconstructed by the joint maximum a posteriori estimation method. The improved infrared thermal radiation model is suitable for scenes with different measurement distances, humidity factors and emissivity. Through multiple experimental verifications of black bodies and high-speed direct-drive fans, this method can significantly reduce measurement errors and improve the accuracy of temperature measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 A flowchart of a method for infrared image temperature calibration provided in one embodiment of the present application.

[0044] Figure 2 A structural connection diagram of a system for infrared image temperature calibration provided in one embodiment of the present application.

[0045] Reference numerals:

[0046] 100 - processing end; 200 - external device; 210 - infrared camera; 220 - parameter collector. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] The present application provides a method for infrared image temperature calibration.

[0049] like Figure 1 As shown, in one embodiment of the present application, a method for infrared image temperature calibration includes:

[0050] S100, determining the atmospheric transmittance.

[0051] S200, establishes an infrared radiation model based on atmospheric transmittance.

[0052] S300, calling the object surface emission first verification information.

[0053] S400, build the forward model.

[0054] S500 forms a joint estimation model of temperature and hyperparameters based on the infrared radiation model, the prior information of the object surface emission and the forward model.

[0055] S600, using an alternating iterative optimization algorithm in a joint estimation model of temperature and hyperparameters to obtain a final temperature result.

[0056] Specifically, the emissivity is regarded as a Gaussian distribution variable with a known mean and variance, and is directly introduced into the forward model for temperature calibration. The advantage of this is that the change in emissivity can be described more flexibly under different measurement conditions, improving the applicability of the model and the accuracy of temperature calibration. By analyzing and correcting the measurement data using the Bayesian inference method, the new model can effectively reduce the measurement error and obtain more accurate temperature measurement results under different measurement distances, humidity and material properties.

[0057] More specifically, the emissivity is no longer considered a fixed value, but is introduced as a random variable with a known mean and variance, so that the changing characteristics of the emissivity are fully considered in the forward model. The model parameters are optimized in combination with Bayesian inference to ensure the accuracy and robustness of temperature measurement in various complex environments. This method makes up for the limitations of emissivity processing in traditional methods, achieves better calibration results in experiments, and proves its reliability in the field of heat source measurement.

[0058] The present embodiment relates to a method for infrared image temperature calibration, which determines the error amount caused by the measured distance between the infrared radiation source and the infrared imager through atmospheric transmittance. The surface emissivity error of the infrared radiation source in the actual measurement process is reduced by calling the prior information of the surface emission of the object. Based on the atmospheric transmittance, an infrared radiation model is established and a forward model is constructed to form a joint estimation model of temperature and hyperparameters. The surface emissivity of the object is derived using the Bayesian inference method, and the prior information of different measurement materials is considered at the same time. Combined with the improved model, the model parameters and temperature distribution are reconstructed by the joint maximum a posteriori estimation method. The improved infrared thermal radiation model is suitable for scenarios with different measurement distances, humidity factors and emissivity. Through multiple experimental verifications of black bodies and high-speed direct-drive fans, this method can significantly reduce measurement errors and improve the accuracy of temperature measurement.

[0059] In one embodiment of the present application, S100 includes:

[0060] S110, obtaining relative humidity, measurement distance, and absorption characteristic data of infrared band.

[0061] S120, relative humidity, measurement distance and absorption characteristic data of infrared band are incorporated into the parameter formula of Beer-Lambert law.

[0062] S130, based on the parameter formula of the Beer-Lambert law, calculates the attenuation data of infrared radiation in the atmosphere.

[0063] S140, using the attenuation data of infrared radiation in the atmosphere, determine an expression for atmospheric transmittance.

[0064] Specifically, the atmospheric transmittance is inferred. In the process of infrared temperature measurement, atmospheric transmittance is a key parameter that directly affects the accuracy of temperature calibration. First, the relative humidity, measurement distance, and absorption characteristics of the infrared band are obtained in the measurement environment. Then, the Beer-Lambert law is used to calculate the attenuation of infrared radiation in the atmosphere.

[0065] The expression of atmospheric transmittance is:

[0066]

[0067] Where τ is the atmospheric transmittance, k(h w ) is the water vapor content h w The relevant attenuation coefficient, d is the propagation distance of infrared radiation, and e is a natural constant.

[0068] Since water vapor has an effect on the attenuation coefficient under different temperature conditions, this embodiment will also consider calculating the relationship between the attenuation coefficient and the temperature and humidity:

[0069] The relationship between the attenuation coefficient and temperature and humidity is:

[0070]

[0071] in, is the temperature T 1 The amount of water vapor condensed, is the temperature T 2 The amount of water vapor condensed, and is the relative humidity at different temperatures, At temperature T 1 The attenuation coefficient when At temperature T 2 The attenuation coefficient when .

[0072] The attenuation coefficient k(h) in the current environment can be determined w ), and the atmospheric transmittance τ is obtained accordingly.

[0073] In one embodiment of the present application, S200 includes:

[0074] S210, define emissivity as a parameter to be attached with a value.

[0075] S220, calling an initial infrared radiation model.

[0076] S230, incorporating the emissivity into the initial infrared radiation model to form an infrared radiation model.

[0077] Specifically, an improved infrared radiation model is established. In a traditional infrared radiation model, the relationship between the measured temperature and the actual temperature is usually limited by the assumption that the emissivity is a constant. However, the present invention regards the emissivity as a variable and directly introduces it into the model. The improved infrared radiation model can more realistically reflect the infrared radiation characteristics under different measurement conditions.

[0078] The expression of the temperature measured by the infrared imager in the infrared radiation model is:

[0079]

[0080] Among them, T mis the temperature measured by the infrared imager, T r is the actual temperature of the target object, ε is the surface emissivity of the object, T u is the background temperature, T a is the atmospheric temperature, and n is a constant related to the infrared sensor material.

[0081] In one embodiment of the present application, S300 includes:

[0082] S310, defining the surface emissivity of the object as a parameter to be attached with a value.

[0083] S320, receiving experimental measurement data of the surface emissivity of the object.

[0084] S330, statistical experimental measurement data.

[0085] S340, using Gaussian distribution, determining the mean and variance of the surface emissivity of the object in the experimental measurement data.

[0086] Specifically, the prior information of the surface emissivity of the object is constructed. The present invention takes the surface emissivity of the object as a variable and constructs its prior information through experimental measurement data. First, multiple experiments are conducted to obtain a set of representative emissivity measurement values ​​on the surface of the same material object and under different environmental conditions through infrared temperature measurement equipment. Then, the measured values ​​obtained from the experiment are statistically analyzed, and it is assumed that the emissivity obeys a Gaussian distribution. By fitting the measured data, the mean value μ of the emissivity is obtained. ε and variance v ε , as the distribution parameter of the emissivity. The Gaussian distribution of the emissivity is expressed as:

[0087] ε~N(ε|μ ε , v ε ) Formula 4

[0088] After constructing the prior information of emissivity, the emissivity model is used for subsequent iterative calculation and temperature calibration. The model can dynamically reflect the changing characteristics of emissivity and ensure the accuracy and robustness of temperature measurement.

[0089] In one embodiment of the present application, S400 includes:

[0090] S410, the forward model of the temperature measured by the infrared imager is expressed as:

[0091] T m =H(T r ,ε)+ζ Formula 5

[0092] Where ζ is the measured swing.

[0093] S420, predetermining a variance of the measured swing amount based on an inverse gamma distribution.

[0094] S430, defining the variance of the actual temperature distribution of the target object as an inverse gamma distribution.

[0095] Specifically, ζ is the measurement uncertainty. In the forward model, the emissivity is introduced as a variable. In this way, the change of emissivity can be dynamically reflected in the temperature measurement process, which improves the flexibility and accuracy of temperature calibration. The scaling operator H(.) is based on the improved infrared radiation model.

[0096] In the forward model, the noise ζ is assumed to be independent and identically distributed and follows a Gaussian distribution, and in fact its variance is unknown and needs to be estimated. Therefore, we regard it as an inverse gamma distribution:

[0097]

[0098] Since the measurement error is assumed to be Gaussian, we have:

[0099] P(T m |T r , v r , v ζ )=N(T m |T r , v ζ ) Formula 7

[0100] And T r is set to be Gaussian and its variance is set to be inverse gamma:

[0101]

[0102] Where T av It is obtained based on prior information and can be calculated according to the improved infrared radiation model.

[0103] In one embodiment of the present application, S500 includes:

[0104] S510, establishing a joint estimation model of temperature and hyperparameters based on the actual temperature of the target object, the surface emissivity of the object, a first hidden variable related to the actual temperature of the target object, and a second hidden variable related to the surface emissivity of the object.

[0105] S520, optimizing the joint estimation model of temperature and hyperparameters using the JMAP estimation method.

[0106] S530: Generate a cost function for a joint estimation model of temperature and hyperparameters.

[0107] Specifically, the expression of the joint estimation model of temperature and hyperparameters is:

[0108] P(Tr ,ε,v r ,v ζ |T m )=P(T m |, T r ,ε,v ξ )P(T r |, T av , v r )

[0109] P(ε|μ ε , ν ε )P(ν r |a t , b t )P(ν ζ |a ζ , b ζ ) Formula 9

[0110] Perform JMAP estimation on the joint estimation model of temperature and hyperparameters to obtain:

[0111]

[0112] More specifically, the alternating minimization cost function is:

[0113]

[0114] The cost function for the joint estimation model of temperature and hyperparameters is:

[0115]

[0116] In one embodiment of the present application, S600 includes:

[0117] Joint posterior probability inference: T r ,ε,v r and v ζ It can be estimated by alternatively minimizing the cost function.

[0118]

[0119] The gradient descent method can be used to obtain The value of .

[0120]

[0121] Similarly, we can update the variable ε to obtain Values:

[0122]

[0123] And v r and v ζBy fixing T r and ε, and v r and v ζ When the derivative of is equal to 0, we calculate:

[0124]

[0125] Use the alternating iterative optimization algorithm, set a predetermined number of iterations, and obtain the final temperature result.

[0126] The present application provides a system for infrared image temperature calibration.

[0127] like Figure 2 As shown, in one embodiment of the present application, a system for infrared image temperature calibration includes a processing end 100 and an external device 200 .

[0128] The processing end 100 is used to execute the infrared image temperature calibration method.

[0129] The external device 200 is connected to the processing end 100 for communication, and the external device 200 includes one or more of an infrared camera 210 and a parameter collector 220 .

[0130] Specifically, the infrared camera 210 is used to acquire infrared images.

[0131] The parameter collector 220 includes a temperature sensor, a humidity sensor, a distance measuring device, etc. The parameter collector 220 is used to measure the atmospheric transmittance and construct a priori information of the surface emission of an object.

[0132] The present embodiment relates to a system for infrared image temperature calibration, in which the processing end 100 calculates the atmospheric transmittance through an external device 200 to determine the error amount caused by the measured distance between the infrared radiation source and the infrared imager. The processing end 100 calls the prior information of the surface emission of the object to reduce the surface emissivity error of the infrared radiation source during the actual measurement process. Based on the atmospheric transmittance, an infrared radiation model is established and a forward model is constructed to form a joint estimation model of temperature and hyperparameters. The surface emissivity of the object is derived using the Bayesian inference method, and the prior information of different measurement materials is considered at the same time. Combined with the improved model, the model parameters and temperature distribution are reconstructed by the joint maximum a posteriori estimation method. The improved infrared thermal radiation model is suitable for scenarios with different measurement distances, humidity factors and emissivity. Through multiple experimental verifications of black bodies and high-speed direct-drive fans, this method can significantly reduce measurement errors and improve the accuracy of temperature measurement.

[0133] The technical features of the above-described embodiments may be arbitrarily combined, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for infrared image temperature calibration, characterized in that: include: Determine atmospheric transmittance; Based on the atmospheric transmittance, an infrared radiation model is established; Call the object surface to emit the first-come-first-served information; Build a forward model; Based on the infrared radiation model, the a priori information of the object surface emission and the forward model, a joint estimation model of temperature and hyperparameters is formed; An alternating iterative optimization algorithm is used in the joint estimation model of temperature and hyperparameters to obtain the final temperature result.

2. The method for infrared image temperature calibration according to claim 1, characterized in that: The determining of the atmospheric transmittance comprises: Obtain relative humidity, measurement distance, and infrared absorption characteristic data; Incorporate relative humidity, measurement distance, and infrared absorption characteristic data into the parameter formula of the Beer-Lambert law; Based on the parameter formula of the Beer-Lambert law, calculate the attenuation data of infrared radiation in the atmosphere; Using the attenuation data of infrared radiation in the atmosphere, the expression of atmospheric transmittance is determined.

3. The method for infrared image temperature calibration according to claim 2, characterized in that: The expression of the atmospheric transmittance is: Where τ is the atmospheric transmittance, k(h w ) is the water vapor content h w The relevant attenuation coefficient, d is the propagation distance of infrared radiation, and e is a natural constant.

4. The method for infrared image temperature calibration according to claim 3, characterized in that: The parameter formula for incorporating relative humidity, measurement distance, and infrared absorption characteristic data into the Beer-Lambert law includes: Determine the relationship between the attenuation coefficient and the temperature and humidity. The relationship between the attenuation coefficient and the temperature and humidity is: in, is the condensation amount of water vapor at temperature T1, is the condensation amount of water vapor at temperature T2, and is the relative humidity at different temperatures, is the attenuation coefficient at temperature T1, is the attenuation coefficient at temperature T2.

5. The method for infrared image temperature calibration according to claim 4, characterized in that: The infrared radiation model is established based on the atmospheric transmittance, including: Define emissivity as a parameter to be assigned; Call the initial infrared radiation model; The emissivity is incorporated into the initial infrared radiation model to form an infrared radiation model.

6. The method for infrared image temperature calibration according to claim 5, characterized in that: The step of incorporating the emissivity into the initial infrared radiation model to form the infrared radiation model includes: An expression for the temperature measured by the infrared imager in the infrared radiation model is defined; the expression for the temperature measured by the infrared imager in the infrared radiation model is: Among them, T m is the temperature measured by the infrared imager, T r is the actual temperature of the target object, ε is the surface emissivity of the object, T u is the background temperature, T a is the atmospheric temperature, and n is a constant related to the infrared sensor material.

7. The method for infrared image temperature calibration according to claim 6, characterized in that: The calling of the object surface emission prior information includes: Define the surface emissivity of the object as a parameter to be assigned; Receive experimental measurement data of the surface emissivity of the object; Statistics experimental measurement data; Using Gaussian distribution, the mean and variance of the surface emissivity of the object in the experimental measurement data are determined.

8. The method for infrared image temperature calibration according to claim 7, characterized in that: The constructing of the forward model comprises: The forward model for determining the temperature measured by the infrared imager is expressed as: T m =H(T r ,e)+g Among them, ζ is the measured swing amount; Based on the inverse gamma distribution, the variance of the measured swing is predetermined; The variance of the actual temperature distribution of the target object is defined as an inverse gamma distribution.

9. The method for infrared image temperature calibration according to claim 8, characterized in that: The method forms a joint estimation model of temperature and hyperparameters based on the infrared radiation model, the a priori information of the object surface emission and the forward model, including: Based on the actual temperature of the target object, the emissivity of the object surface, a first hidden variable related to the actual temperature of the target object, and a second hidden variable related to the emissivity of the object surface, a joint estimation model of temperature and hyperparameters is established; Use JMAP estimation method to optimize the joint estimation model of temperature and hyperparameters; Generates a cost function for the joint estimation model of temperature and hyperparameters.

10. A system for infrared image temperature calibration, characterized in that: include: A processing end, used to execute the infrared image temperature calibration method according to any one of claims 1 to 9; An external device is communicatively connected to the processing end, and the external device includes one or more of an infrared camera and a parameter collector.

Images