CN-121977703-A - Infrared radiation calibration method, device, equipment and storage medium considering test distance

CN121977703ACN 121977703 ACN121977703 ACN 121977703ACN-121977703-A

Abstract

The invention discloses an infrared radiation calibration method, a device, equipment and a storage medium considering a test distance, wherein the method comprises the steps of constructing an infrared radiation physical model considering the integration time of an infrared detector and the test distance of a target; fitting parameters in the infrared radiation physical model by utilizing the training set and the coefficient learning network model to obtain a parameter optimal solution, and applying the optimal solution to the infrared radiation physical model to finish infrared radiation calibration. According to the method, the influence of the test distance and the integration time on the relation between the gray level image and the radiation brightness is considered, so that an infrared radiation physical model is constructed, the accuracy of a radiation calibration result is improved, the gray level images under different test distance and integration time combinations and the parameters of the infrared radiation physical model are optimized by a training set formed by the blackbody radiation brightness corresponding to each gray level image, so that the infrared radiation physical model is provided with an optimal solution, and the measurement accuracy of an infrared radiation measurement system on the target radiation brightness distribution is improved.

Inventors

ZHOU HUIXIN
YAN XUN
WANG BINGJIAN
LAI RUI
QIN HANLIN
SHI JIN
XIANG PEI
SONG JIANGLUQI
QI SHUXIA
TENG XIANG
SUN RUIYANG
PU ZHENG
Xu Xiujin

Assignees

西安电子科技大学

Dates

Publication Date: 20260505
Application Date: 20251231

Claims (10)

1. An infrared radiation calibration method taking test distance into consideration, the method comprising: Constructing an infrared radiation physical model considering the integration time of an infrared detector and the test distance of a target to be tested, wherein the infrared radiation physical model is used for establishing the relation between the radiation brightness and the image gray value output by the infrared detector; Acquiring a training set, wherein the training set comprises gray images of the infrared detector under different integration time and blackbody test distance combinations and blackbody radiation brightness corresponding to each gray image; And fitting parameters in the infrared radiation physical model by utilizing the training set and a preset coefficient learning network model to obtain an optimal solution of the parameters in the infrared radiation physical model, so that the optimal solution of the parameters in the infrared radiation physical model is substituted into the infrared radiation physical model, and infrared radiation calibration is completed.
2. The method according to claim 1, wherein the constructing an infrared radiation physical model taking into account an integration time of the infrared detector and a test distance of the object to be measured, specifically comprises: According to different integration time of the infrared detector, response of the infrared detector to the radiation brightness of the incident target to be detected and radiometric calibration bias under the test distance of the target to be detected, an initial relation between the radiation brightness and a gray level image output by the infrared detector is constructed; Constructing an expression of the absorption coefficient changing along with the test distance according to the initial absorption coefficient when the test distance between the infrared detector and the target is zero and the absorption coefficient component changing along with the test distance; based on the beer-lambert law, correcting the initial relation according to the expression of the absorption coefficient to obtain a target relation between corrected radiation brightness and a gray level image output by the infrared detector, wherein the infrared radiation physical model comprises the target relation between the radiation brightness and the gray level image output by the infrared detector.
3. The method of claim 2, wherein the expression of the absorption coefficient as a function of the test distance is represented by: In the formula, To test the absorption coefficient at distance d, An initial absorption coefficient for a test distance of zero; And A first absorption coefficient component and a second absorption coefficient component that vary with the test distance, The linear change in absorption with distance is characterized, The nonlinear change in absorption with distance is characterized.
4. The method of claim 2, wherein when the object to be measured is a black body, an initial relation between a black body radiation brightness and a gray image output from the infrared detector is expressed as follows: In the formula, When the black body temperature is T b and the integration time is T, the gray value output by the infrared detector; K is the response rate of the infrared detector to the radiation brightness of the incident blackbody; Is of blackbody temperature The radiation brightness at that time; Bias radiation induced for ambient temperature; Bias radiation induced by factors internal to the infrared detector.
5. The method as claimed in claim 4, wherein the target relation between the corrected radiance and the grayscale image output by the infrared detector is expressed as: In the formula, Is of blackbody temperature The radiation brightness of the black body; Is of blackbody temperature When the integration time is t, the gray value output by the infrared detector; Bias radiation induced for ambient temperature; k is the response rate of the infrared detector to the radiation brightness of an incident blackbody; An initial absorption coefficient for a test distance d of zero; And A first absorption coefficient component and a second absorption coefficient component, respectively, that vary with the test distance.
6. The method of claim 1, wherein the learning a network model using the training set and a preset coefficient, fitting parameters in the infrared radiation physical model, and obtaining an optimal solution of the parameters in the infrared radiation physical model, specifically includes: Initializing parameters in the infrared radiation physical model, and presetting an initial value of each parameter; Inputting sample data of a target sample into an initialized infrared radiation physical model to obtain a predicted value of blackbody radiation brightness output by the infrared radiation physical model, wherein the target sample is any one of all samples in the training set, and at least comprises an integration time of the infrared detector, a blackbody test distance, a gray value of the blackbody output by the infrared detector under the integration time and the blackbody test distance, and an actual value of blackbody radiation brightness; calculating based on a loss function, a predicted value and an actual value of blackbody radiation brightness in the coefficient learning network model to obtain the loss of the coefficient learning network model and the gradient of the loss to the predicted value; And based on a gradient descent method, carrying out parameter updating on the gradient of the predicted value according to the partial derivative of each parameter and the loss until the loss meets a preset stopping iteration condition so as to determine the optimal solution of the parameter in the infrared radiation physical model.
7. The method as recited in claim 6, wherein the parameters to be optimized in the infrared radiation physical model include an initial absorption coefficient when the infrared detector test distance is zero, a first absorption coefficient component and a second absorption coefficient component which change with the test distance, an ambient temperature induced radiation bias, an infrared detector internal factor induced radiation bias, and a response rate of the infrared detector to incident blackbody radiation brightness; The parameter update formula is: In the formula, An initial absorption coefficient for a test distance of zero; And A first absorption coefficient component and a second absorption coefficient component which are respectively changed along with the test distance; Bias radiation induced for ambient temperature; k is the response rate of the infrared detector to the radiation brightness of an incident blackbody; is the learning rate; Is loss; Is a predicted value.
8. An infrared radiation calibration device taking into account a test distance, the device comprising: the infrared radiation detection device comprises a model construction unit, a detection unit and a detection unit, wherein the model construction unit is used for constructing an infrared radiation physical model considering the integration time of an infrared detector and the test distance of a target to be detected, and the infrared radiation physical model is used for establishing the relation between the radiation brightness and the image gray value output by the infrared detector; The data acquisition unit is used for acquiring a training set, wherein the training set comprises gray images of the infrared detector under different integration time and blackbody test distance combinations and blackbody radiation brightness corresponding to each gray image; And the parameter optimizing unit is used for utilizing the training set and a preset coefficient to learn a network model, fitting the parameters in the infrared radiation physical model to obtain an optimal solution of the parameters in the infrared radiation physical model, and substituting the optimal solution of the parameters in the infrared radiation physical model into the infrared radiation physical model to finish infrared radiation calibration.
9. A computer readable storage medium storing a computer program, which when executed by a processor causes the processor to perform the steps of the method according to any one of claims 1 to 7.
10. A computer device comprising a memory and a processor, wherein the memory stores a computer program which, when executed by the processor, causes the processor to perform the steps of the method of any of claims 1 to 7.

Description

Infrared radiation calibration method, device, equipment and storage medium considering test distance Technical Field The present invention relates to the field of optical and photonics technologies, and in particular, to an infrared radiation calibration method, an infrared radiation calibration device, an infrared radiation calibration apparatus, and a storage medium considering a test distance. Background An infrared radiation measurement system receives a radiation signal of a target through non-contact detection and converts the received radiation signal into an electrical signal. The electrical signals are then digitized to output an infrared image. And (3) finishing target infrared radiation characteristic measurement by mapping the gray value and the radiation brightness of the obtained target image, and finally obtaining the radiation brightness distribution and the temperature distribution of the target. The obtained relevant information may provide a priori features for subsequent target detection, identification, and segmentation. The radiation calibration of an infrared radiation measurement system is a necessary prerequisite for achieving a measurement of the infrared radiation characteristics of the target. Radiometric calibration is achieved by determining the responsivity and bias of an infrared radiation measurement system to target input radiation. Firstly, a model between the radiation brightness of the black body and the output gray value of the infrared radiation measurement system is established. And then measuring the target to be measured according to the model. In the actual measurement scene, the existing radiation calibration method has the defects of insufficient measurement precision because the test distance is not fixed and the adopted radiation calibration model does not consider the change of the test distance, so that the measurement result of the current infrared radiation measurement system on the target radiation brightness distribution and the temperature distribution is inaccurate. Disclosure of Invention Based on this, it is necessary to provide an infrared radiation calibration method, an infrared radiation calibration device, and a storage medium, which consider the test distance, so as to improve the accuracy of the radiation calibration result and the measurement accuracy of the infrared radiation measurement system on the target radiance distribution. To achieve the above object, a first aspect of the present application provides an infrared radiation calibration method considering a test distance, the method comprising: Constructing an infrared radiation physical model considering the integration time of an infrared detector and the test distance of a target to be tested, wherein the infrared radiation physical model is used for establishing the relation between the radiation brightness and the image gray value output by the infrared detector; Acquiring a training set, wherein the training set comprises gray images of the infrared detector under different integration time and blackbody test distance combinations and blackbody radiation brightness corresponding to each gray image; And fitting parameters in the infrared radiation physical model by utilizing the training set and a preset coefficient learning network model to obtain an optimal solution of the parameters in the infrared radiation physical model, so that the optimal solution of the parameters in the infrared radiation physical model is substituted into the infrared radiation physical model, and infrared radiation calibration is completed. Further, the construction of the infrared radiation physical model taking the integration time of the infrared detector and the test distance of the target to be tested into consideration specifically comprises the following steps: According to different integration time of the infrared detector, response of the infrared detector to the radiation brightness of the incident target to be detected and radiometric calibration bias under the test distance of the target to be detected, an initial relation between the radiation brightness and a gray level image output by the infrared detector is constructed; Constructing an expression of the absorption coefficient changing along with the test distance according to the initial absorption coefficient when the test distance between the infrared detector and the target is zero and the absorption coefficient component changing along with the test distance; based on the beer-lambert law, correcting the initial relation according to the expression of the absorption coefficient to obtain a target relation between corrected radiation brightness and a gray level image output by the infrared detector, wherein the infrared radiation physical model comprises the target relation between the radiation brightness and the gray level image output by the infrared detector. Further, the expression of the absorption coefficient according to th