Temperature correction method of thermal infrared imager

By training the ambient temperature prediction model in the infrared thermal image camera and adjusting the infrared image temperature, the problem of deviation of the ambient temperature prediction result in the existing technology is solved, and more accurate and fast ambient temperature prediction is achieved, and the temperature measurement accuracy of infrared equipment is improved. .

CN118999801AActive Publication Date: 2024-11-22HEILONGJIANG ELECTRIC POWER SCIENCE RESEARCH INSTITUTE +3
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411099779.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-22
Estimated Expiration
2044-08-12

AI Technical Summary

Technical Problem

When the existing infrared imaging equipment faces the temperature change of the external environment, the ambient temperature prediction results form a closed loop due to the lack of feedback, resulting in a large deviation in the temperature measurement accuracy. Especially when the internal temperature sensor of the equipment is not updated in time, the prediction results have a large deviation. .

Method used

The temperature correction method of infrared thermal imager is adopted to train the ring temperature prediction model by sampling temperature within the set ambient temperature range, and the ambient temperature prediction model is predicted using the model trained by sample data, and the difference between the temperature in the background area and the predicted value is judged. to adjust the infrared image temperature to ensure the accuracy of the temperature measurement results.

Benefits of technology

It achieves more accurate and fast ambient temperature prediction, improves the temperature measurement accuracy of infrared equipment in harsh environments, avoids the limitations of traditional methods in low temperature environments, and ensures the closed-loop accuracy of ambient temperature prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118999801A_ABST
    Figure CN118999801A_ABST
Patent Text Reader

Abstract

The invention discloses a temperature correction method of a thermal infrared imager, and belongs to the field of infrared temperature measurement. The invention aims to solve the problem that the current environment temperature prediction adopts an open-loop mode, so that the environment temperature prediction result has great deviation. Obtaining an infrared image of the current measured object and temperature data of each component on the current thermal infrared imager, and inputting the temperature data of each component on the thermal infrared imager to the trained environment temperature prediction to obtain an environment temperature prediction value; obtaining an infrared image of the current measured object; whether the absolute value of the difference value between the background area temperature in the infrared image of the current measured object and the predicted value of the environment temperature is smaller than a preset value or not is judged, if yes, the infrared image temperature of the current measured object serves as a final result to be output, and if not, the infrared image temperature of the current measured object serves as a final result. And replacing the background area temperature in the infrared image temperature of the current measured object with an environment temperature prediction value, and outputting the obtained foreground area temperature and the replaced infrared image background temperature as the infrared image temperature of the measured object. The method is used for correcting the temperature of the thermal infrared imager.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a temperature correction method and belongs to the field of infrared temperature measurement. Background Art

[0002] As infrared thermal imaging technology becomes more mature, users have higher and higher requirements for the temperature measurement accuracy of infrared imaging equipment such as infrared thermal imagers. The temperature measurement accuracy of infrared imaging equipment is affected by many external factors, such as emissivity, ambient temperature, atmospheric attenuation, and the distance between the object being measured and the temperature measurement equipment. Among them, ambient temperature is a more important factor, and it is generally necessary to use ambient temperature to correct the temperature measurement results.

[0003] The current ambient temperature prediction methods are all open-loop strategies, that is, one or more of the many sensors such as the device's own temperature sensor, external temperature and humidity sensors, and wind speed sensors are used to predict the current real ambient temperature, thereby correcting the temperature measurement results. Since the current ambient temperature prediction uses an open-loop method and lacks feedback to form a closed loop, especially when the external ambient temperature changes significantly and the temperature sensors inside the device have not been updated in time, the ambient temperature prediction results will have a large deviation. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the current ambient temperature prediction uses an open-loop method. When the external ambient temperature changes greatly and the temperature sensors inside the equipment cannot be updated in time, the ambient temperature prediction results will have a large deviation. A temperature correction method for an infrared thermal imager is proposed.

[0005] A temperature correction method for an infrared thermal imager, the method comprising the following steps:

[0006] Step 1, uniformly select n sampling temperatures from a set ambient temperature range, place the infrared thermal imager in a temperature box, set the temperature in the temperature box to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, perform data processing on the actual temperature data of each component, and obtain the processed actual temperature data of each component;

[0007] The processed actual temperature data of each component and the corresponding set sampling temperature are taken as a sample; the processed temperature data of each component in the sample are taken as the input of the ambient temperature prediction model, and the set sampling temperature is taken as the output of the ambient temperature prediction model;

[0008] Step 2: Use the samples to train the ambient temperature prediction model to obtain a trained ambient temperature prediction model;

[0009] Step 3: Obtain the infrared image of the current object being measured and the temperature data of each component on the current infrared thermal imager, input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction, and obtain the ambient temperature prediction value;

[0010] Step 4: acquiring an infrared image of the current object under test, wherein the infrared image of the current object under test includes a foreground area and a background area;

[0011] Step 5: determine whether the absolute value of the difference between the background area temperature in the infrared image of the current object being measured and the predicted value of the ambient temperature is less than a preset value. If so, output the infrared image temperature of the current object being measured as the final result. If not, execute step 6.

[0012] Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted value of the ambient temperature, and output the foreground area temperature obtained in step 4 and the replaced infrared image background temperature as the infrared image temperature of the object under test.

[0013] Preferably, in step 1, the temperature data of each component on the infrared thermal imager includes: device focal plane temperature, device shutter temperature, device cavity temperature, device focal plane temperature rise rate, device shutter temperature rise rate and device cavity temperature rise rate.

[0014] Preferably, the specific process of step 6 is:

[0015] Acquire background radiation energy of the background area and foreground radiation energy of the foreground area;

[0016] The predicted ambient temperature is converted into reflected radiation according to the blackbody radiation law;

[0017] Replace the reflected radiation amount with the background radiation energy;

[0018] The foreground radiation energy of the foreground area and the replaced background radiation energy are brought into the calibrated radiation-temperature mapping curve in the infrared thermal imager to obtain the background temperature of the infrared image.

[0019] Preferably, the specific process of obtaining the background radiation energy of the background area is:

[0020] Perform radiation histogram statistics on the background area;

[0021] Convolve the radiation amount histogram to obtain the radiation amount range of the background area with the highest confidence;

[0022] The average of the background area radiation range with the highest confidence level is taken as the background radiation energy.

[0023] Preferably, ambient temperature prediction is implemented using a multiple linear regression model.

[0024] Preferably, in step 1, the temperature data of each component is processed to obtain the processed temperature data of each component. The specific process is:

[0025] The temperature data of each component is downsampled to obtain the downsampled data, each downsampled data is projected into a space, the projections of each downsampled data on multiple vectors in the space are multiplied to obtain each product result, each product result is weighted, and the weighted data are combined into the temperature data of each component.

[0026] The beneficial effects of the present invention are:

[0027] The present invention uses an environmental prediction model to predict the ambient temperature, while the existing method uses sensors to collect the ambient temperature. Compared with the existing method, the present invention obtains the ambient temperature more accurately and faster, and the existing environmental sensors cannot collect ambient temperatures below zero degrees, which has limitations.

[0028] The present invention determines whether the absolute value of the difference between the background area temperature in the infrared image of the current object under test and the predicted value of the ambient temperature is less than a preset value. If so, the infrared image temperature of the current object under test is output as the final result; if not, it indicates that there is a deviation in the result of the ambient temperature prediction value, and the background area temperature in the infrared image temperature of the current object under test is replaced with the predicted value of the ambient temperature, and the foreground area temperature and the replaced infrared image background temperature are output as the infrared image temperature of the object under test. The present invention ensures the accuracy of the ambient temperature prediction through a closed-loop ambient temperature prediction method, thereby improving the temperature measurement accuracy of infrared equipment in harsh environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The figure is a flow chart of the temperature correction method of the infrared thermal imager. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0032] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0033] Example:

[0034] Combination Figure 1 The present embodiment describes a temperature calibration method for an infrared thermal imager, the method comprising the following steps:

[0035] Step 1, uniformly select n sampling temperatures from a set ambient temperature range, place the infrared thermal imager in a temperature box, set the temperature in the temperature box to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, perform data processing on the actual temperature data of each component, and obtain the processed actual temperature data of each component;

[0036] The processed actual temperature data of each component and the corresponding set sampling temperature are taken as a sample; the processed temperature data of each component in the sample are taken as the input of the ambient temperature prediction model, and the set sampling temperature is taken as the output of the ambient temperature prediction model;

[0037] Step 2: Use the samples to train the ambient temperature prediction model to obtain a trained ambient temperature prediction model;

[0038] Step 3: Obtain the infrared image of the current object being measured and the temperature data of each component on the current infrared thermal imager, input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction, and obtain the ambient temperature prediction value;

[0039] Step 4: acquiring an infrared image of the current object under test, wherein the infrared image of the current object under test includes a foreground area and a background area;

[0040] Step 5: determine whether the absolute value of the difference between the background area temperature in the infrared image of the current object being measured and the predicted value of the ambient temperature is less than a preset value. If so, output the infrared image temperature of the current object being measured as the final result. If not, execute step 6.

[0041] Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted value of the ambient temperature, and output the foreground area temperature obtained in step 4 and the replaced infrared image background temperature as the infrared image temperature of the object under test.

[0042] Specifically, in step 1, the set ambient temperature range refers to multiple real ambient temperatures. Assume that the operating temperature range defined by the current device is T min ~T max , n ambient temperature sampling points are uniformly selected during the period (in this embodiment ), corresponding to each ambient temperature sampling point, place the device in the corresponding ambient temperature incubator, and record all temperature data from the device startup to the thermal state stabilization process. One data contains the following information (the temperature rise rate is defined as the temperature rise of the corresponding sensor within 5 minutes): the current real ambient temperature T amb 、Current device focal plane temperature T jiao、Current device shutter temperature T s 、Current device cavity temperature T j , Current device focal plane temperature rise rate S jiao , Current device shutter temperature rise rate S s , the current temperature rise rate S in the equipment cavity j One piece of data constitutes a sample, which eventually forms a training data set containing multiple samples.

[0043] In step 3, the infrared image and the infrared thermal imager temperature data are acquired in real time, and the time of acquisition of the two is basically the same. Among them, the infrared image is acquired through the infrared detection module inside the infrared thermal imager, and the infrared thermal imager temperature data is acquired through the temperature sensor inside the infrared thermal imager. The temperature sensor includes but is not limited to the focal plane temperature sensor, the shutter temperature sensor, and the cavity temperature sensor. The corresponding infrared device temperature data selects some temperature parameters associated with the ambient temperature of the device. In this embodiment, the device temperature data includes the device focal plane temperature, the device shutter temperature, the device cavity temperature, the device focal plane temperature rise rate, the device shutter temperature rise rate, and the device cavity temperature rise rate, wherein the device focal plane temperature rise rate, the device shutter temperature rise rate, and the device cavity temperature rise rate are calculated based on the temperature rise of the corresponding sensor per unit time. In other embodiments, only some of the above parameters may be included, and other related parameters may also be included. To ensure the prediction accuracy, the parameters in the device temperature data include at least two types.

[0044] The above ambient temperature prediction value is only obtained through open-loop prediction of the ambient temperature prediction model. When the external ambient temperature changes greatly and the temperature sensors inside the equipment have not been updated in time, the ambient temperature prediction result will have a large deviation, so the accuracy of the prediction result needs to be further verified. Since the background area temperature in the infrared image of the current object being measured should be basically consistent with the ambient temperature prediction value, the accuracy of the infrared image of the current object being measured can be verified by judging whether the two are consistent.

[0045] In some embodiments, it is determined whether the temperature of the background area in the infrared image of the current object being measured is consistent with the predicted value of the ambient temperature. Specifically, it can be determined whether the difference between the temperature of the background area in the infrared image of the current object being measured and the predicted value of the ambient temperature is less than a certain threshold. For example, if The ambient temperature prediction is considered correct, and the temperature matrix of the entire image is output based on the predicted ambient temperature; otherwise, the current prediction is considered inaccurate and the prediction value needs to be readjusted until it meets After this condition, the temperature matrix of the entire image is output.

[0046] The ambient temperature prediction module is only turned on when the device temperature data is not completely stable, and the ambient temperature prediction model is used to predict the ambient temperature to obtain the current ambient temperature prediction value; when the device temperature data is completely stable, the ambient temperature prediction value output by the ambient temperature prediction model is basically unchanged. At this time, the ambient temperature prediction module can be turned off, and the ambient temperature prediction value obtained by the ambient temperature prediction model for the last time is used as the current ambient temperature prediction value to reduce the device running time and memory. The definition of a parameter of the device temperature data being stable can be that the range of the corresponding temperature sensor within a certain period of time is less than a fixed threshold value θ. Taking the current shutter temperature as an example, the shutter temperature transmission value within 5 minutes is When the ambient temperature prediction module is turned on, the value of θ in this embodiment is 0.5°C.

[0047] The following introduces the composition of the temperature data of each component on the infrared thermal imager: The temperature data of each component on the infrared thermal imager includes: device focal plane temperature, device shutter temperature, device cavity temperature, device focal plane temperature rise rate, device shutter temperature rise rate and device cavity temperature rise rate.

[0048] The following further defines that the specific process of step 6 is:

[0049] Acquire background radiation energy of the background area and foreground radiation energy of the foreground area;

[0050] The predicted ambient temperature is converted into reflected radiation according to the blackbody radiation law;

[0051] Replace the reflected radiation amount with the background radiation energy;

[0052] The foreground radiation energy of the foreground area and the replaced background radiation energy are brought into the calibrated radiation-temperature mapping curve in the infrared thermal imager to obtain the background temperature of the infrared image.

[0053] Specifically, the infrared image is segmented into foreground and background to obtain the background area. There are many specific segmentation methods, such as threshold segmentation, edge detection, etc., which will not be described in detail here. Then, the background radiation energy is obtained according to the segmented background area. Specifically, the radiation amount histogram statistics of the segmented background area is first performed:

[0054]

[0055] Among them, pixel(x,y) is the radiation intensity of the pixel at the coordinate point (x,y), width and height are the image width and height respectively, and hist(i) is the number of pixels with radiation intensity i.

[0056] The radiation histogram is convolved with a one-dimensional template with a bandwidth of 50 and coefficients all equal to 1 to obtain the radiation range of the background area with the highest confidence. The mean of the radiation range of the background area with the highest confidence is taken as the background radiation energy Ybg The background radiation energy Y calculated in this way bg It has a high confidence level, making it easier to obtain accurate background temperature later.

[0057] The radiation-temperature mapping curve is calibrated in advance and stored in the device. The radiation-temperature mapping curve can be used to directly calculate the background radiation Y b ' g Get the corresponding infrared image background temperature T bg .

[0058] Of course, the background radiation energy Y bg It is not only affected by the ambient temperature, but also by other factors such as the distance between the object being measured and the temperature measuring device. In this embodiment, only the ambient temperature is used to determine the background radiation energy Y. bg Correction is performed to obtain the true background emission radiation Y b ' g In other embodiments, other factors may be considered to make corrections, and the correction methods are various and not limited to the method provided in this embodiment.

[0059] The specific process of obtaining background radiation energy is further defined below:

[0060] Perform radiation histogram statistics on the background area;

[0061] Convolve the radiation amount histogram to obtain the radiation amount range of the background area with the highest confidence;

[0062] The average of the background area radiation range with the highest confidence level is taken as the background radiation energy.

[0063] The following introduces the optimization method of the model: the model is implemented using a multiple linear regression model.

[0064] A multivariate linear regression model is established with the temperature data of each component on the infrared thermal imager as input and the ambient temperature prediction value as output, that is, the following linear model is constructed:

[0065] T amb =β1*T jiao +β2*T s +β3*T j +β4*S jiao +β5*S s +β6*S j +β7

[0066] In the above formula, β1~β7 are model parameters;

[0067] The model is trained using the above training set, that is, for any sample i in the training set, the following relationship exists:

[0068]

[0069] The loss function is defined as:

[0070]

[0071] In the above formula, is the predicted value of the ambient temperature in the i-th sample.

[0072] The loss function is converged by the gradient descent method, and finally the model parameters β1~β7 are obtained, that is, the trained multivariate linear regression model is obtained and used as the ambient temperature prediction model.

[0073] The following describes how to process the temperature data of each component:

[0074] The temperature data of each component is downsampled to obtain the downsampled data, each downsampled data is projected into a space, the projections of each downsampled data on multiple vectors in the space are multiplied to obtain each product result, the product results greater than a preset value are deleted, and the downsampled data corresponding to each remaining product result are combined into the temperature data of each component.

[0075] Specifically, the temperature data of each component needs to be processed, otherwise the environmental prediction model cannot converge. The specific processing method is to downsample and then reorganize and break it up according to the format required for training.

[0076] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. It should therefore be understood that many modifications may be made to the exemplary embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in a manner different from that described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in other described embodiments.

Claims

1. A temperature correction method for an infrared thermal imager, characterized in that: The method comprises the following steps: Step 1, uniformly select n sampling temperatures from a set ambient temperature range, place the infrared thermal imager in a temperature box, set the temperature in the temperature box to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, perform data processing on the actual temperature data of each component, and obtain the processed actual temperature data of each component; The processed actual temperature data of each component and the corresponding set sampling temperature are taken as a sample; the processed temperature data of each component in the sample are taken as the input of the ambient temperature prediction model, and the set sampling temperature is taken as the output of the ambient temperature prediction model; Step 2: Use the samples to train the ambient temperature prediction model to obtain a trained ambient temperature prediction model; Step 3: Obtain the infrared image of the current object being measured and the temperature data of each component on the current infrared thermal imager, input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction, and obtain the ambient temperature prediction value; Step 4: acquiring an infrared image of the current object under test, wherein the infrared image of the current object under test includes a foreground area and a background area; Step 5: determine whether the absolute value of the difference between the background area temperature in the infrared image of the current object being measured and the predicted value of the ambient temperature is less than a preset value. If so, output the infrared image temperature of the current object being measured as the final result. If not, execute step 6. Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted value of the ambient temperature, and output the foreground area temperature obtained in step 4 and the replaced infrared image background temperature as the infrared image temperature of the object under test.

2. The temperature correction method of the infrared thermal imager according to claim 1, characterized in that: In step 1, the temperature data of each component on the infrared thermal imager includes: device focal plane temperature, device shutter temperature, device cavity temperature, device focal plane temperature rise rate, device shutter temperature rise rate and device cavity temperature rise rate.

3. The temperature correction method of the infrared thermal imager according to claim 1, characterized in that: The specific process of step 6 is as follows: Acquire background radiation energy of the background area and foreground radiation energy of the foreground area; The predicted ambient temperature is converted into reflected radiation according to the blackbody radiation law; Replace the reflected radiation amount with the background radiation energy; The foreground radiation energy of the foreground area and the replaced background radiation energy are brought into the calibrated radiation-temperature mapping curve in the infrared thermal imager to obtain the background temperature of the infrared image.

4. The temperature correction method of the infrared thermal imager according to claim 3, characterized in that: The specific process of obtaining the background radiation energy of the background area: Perform radiation histogram statistics on the background area; Convolve the radiation amount histogram to obtain the radiation amount range of the background area with the highest confidence; The average of the background area radiation range with the highest confidence level is taken as the background radiation energy.

5. The temperature correction method of the infrared thermal imager according to claim 1, characterized in that: The ambient temperature prediction is realized by using the multivariate linear regression model.

6. The temperature correction method of the infrared thermal imager according to claim 1, characterized in that: In step 1, the temperature data of each component is processed to obtain the processed temperature data of each component. The specific process is as follows: The temperature data of each component is downsampled to obtain the downsampled data, each downsampled data is projected into a space, the projections of each downsampled data on multiple vectors in the space are multiplied to obtain each product result, each product result is weighted, and the weighted data are combined into the temperature data of each component.

Citation Information

Patent Citations

  • Zero shutter thermal infrared imager based on voice operated exchange (VOX) detector and use method thereof

    CN103162843A

  • Detection system of infrared thermal imaging equipment

    CN110530529A

  • Thermal imaging system capable of improving imaging quality and temperature measurement precision

    CN202024821U

  • Temperature measurement value processing device, heat generation body detection device, temperature measurement value processing method, and program

    EP4206632A1

  • Apparatus and method for validating leak survey results

    US20160238451A1