Temperature correction method for infrared thermal imager

Abstract

The application relates to a temperature correction method of an infrared thermal imager and belongs to the field of infrared temperature measurement. In order to solve the problem that current ring temperature prediction uses an open loop mode, resulting in a large deviation of the ring temperature prediction result, the application obtains the infrared image of a current measured object and the temperature data of each component on the infrared thermal imager, inputs the temperature data of each component on the infrared thermal imager into a trained ring temperature prediction, and obtains an environment temperature prediction value; the infrared image of the current measured object is obtained; whether the absolute value of the difference between the background area temperature in the infrared image of the current measured object and the environment temperature prediction value is less than a preset value is judged; if yes, the infrared image temperature of the current measured object is taken as the final result output; and if not, the background area temperature in the infrared image temperature of the current measured object is replaced by the environment temperature prediction value, and the obtained foreground area temperature and the replaced infrared image background temperature are taken as the infrared image temperature of the measured object and output. The application is used for correcting the temperature of the infrared thermal imager.

Description

Technical Field

[0001] This invention relates to temperature correction methods and belongs to the field of infrared thermometry. Background Technology

[0002] As infrared thermal imaging technology matures, users are demanding higher accuracy from infrared imaging devices such as infrared thermal imagers. However, the accuracy of infrared imaging is affected by many external factors, such as emissivity, ambient temperature, atmospheric attenuation, and the distance between the object being measured and the measuring device. Ambient temperature is a particularly important factor, and it is generally necessary to use ambient temperature data to correct for the measurement results.

[0003] Current ambient temperature prediction methods all employ open-loop strategies. This means they rely on one or more sensors, such as the device's own temperature sensor, external temperature and humidity sensors, and wind speed sensors, to predict the actual ambient temperature and then correct the readings. Because current ambient temperature prediction methods use an open-loop approach, they lack feedback to form a closed loop. This leads to significant deviations in the predicted ambient temperature, especially when there are large changes in external ambient temperature before the internal temperature sensors have updated in time. Summary of the Invention

[0004] The purpose of this invention is to address the problem that current ambient temperature prediction methods all use an open-loop approach, which leads to significant deviations in the prediction results when the external ambient temperature changes drastically and the internal temperature sensors of the equipment fail to update in time. This invention proposes a temperature correction method for infrared thermal imagers.

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

[0006] Step 1: Select n sampling temperatures uniformly from the set ambient temperature range, place the infrared thermal imager in the temperature chamber, set the temperature in the temperature chamber to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, process the actual temperature data of each component, and obtain the processed actual temperature data of each component.

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

[0008] Step 2: Train the ambient temperature prediction model using samples to obtain the trained ambient temperature prediction model;

[0009] Step 3: Acquire the infrared image of the object being measured and the temperature data of each component on the infrared thermal imager. Input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction system to obtain the predicted ambient temperature value.

[0010] Step 4: Acquire the infrared image of the object being measured, wherein the infrared image of the object being measured includes a foreground region and a background region;

[0011] Step 5: Determine whether the absolute value of the difference between the background area temperature and the predicted ambient temperature in the infrared image of the current object being measured is less than the preset value. If yes, output the infrared image temperature of the current object being measured as the final result. If no, proceed to step 6.

[0012] Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted ambient temperature value, 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: the focal plane temperature, the shutter temperature, the cavity temperature, the focal plane temperature rise rate, the shutter temperature rise rate, and the cavity temperature rise rate.

[0014] Preferably, step 6 is performed as follows:

[0015] Obtain the background radiation energy of the background region and the foreground radiation energy of the foreground region;

[0016] Based on the blackbody radiation law, the predicted ambient temperature is converted into reflected radiation.

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

[0018] The foreground radiation energy of the foreground region and the replaced background radiation energy are fed 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 for obtaining the background radiation energy of the background region is as follows:

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

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

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

[0023] Preferably, ambient temperature prediction is achieved 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 as follows:

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

[0026] The beneficial effects of this invention are:

[0027] This invention uses an environmental prediction model to predict ambient temperature, while existing methods use sensors to collect ambient temperature. Compared with existing methods, this invention obtains ambient temperature more accurately and quickly. Furthermore, existing environmental sensors have limitations in collecting ambient temperatures below zero degrees Celsius.

[0028] This invention determines whether the absolute value of the difference between the background region temperature in the infrared image of the current measured object and the predicted ambient temperature is less than a preset value. If so, the infrared image temperature of the current measured object is output as the final result. If not, it indicates that the predicted ambient temperature result has a deviation. In this case, the background region temperature in the infrared image temperature of the current measured object is replaced with the predicted ambient temperature, and the foreground region temperature and the replaced infrared image background temperature are output as the infrared image temperature of the measured object. This invention uses a closed-loop ambient temperature prediction method to ensure the accuracy of ambient temperature prediction, thereby improving the temperature measurement accuracy of infrared devices in harsh environments. Attached Figure Description

[0029] Figure 1 This is a flowchart of the temperature correction method for infrared thermal imagers. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0032] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0033] Example:

[0034] Combination Figure 1 This embodiment describes a temperature correction method for an infrared thermal imager, which includes the following steps:

[0035] Step 1: Select n sampling temperatures uniformly from the set ambient temperature range, place the infrared thermal imager in the temperature chamber, set the temperature in the temperature chamber to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, process the actual temperature data of each component, and obtain the processed actual temperature data of each component.

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

[0037] Step 2: Train the ambient temperature prediction model using samples to obtain the trained ambient temperature prediction model;

[0038] Step 3: Acquire the infrared image of the object being measured and the temperature data of each component on the infrared thermal imager. Input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction system to obtain the predicted ambient temperature value.

[0039] Step 4: Acquire the infrared image of the object being measured, wherein the infrared image of the object being measured includes a foreground region and a background region;

[0040] Step 5: Determine whether the absolute value of the difference between the background area temperature and the predicted ambient temperature in the infrared image of the current object being measured is less than the preset value. If yes, output the infrared image temperature of the current object being measured as the final result. If no, proceed to step 6.

[0041] Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted ambient temperature value, 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 actual ambient temperatures. Assume the current device's defined operating temperature range is T. min ~T max In this embodiment, n ambient temperature sampling points are evenly selected. For each ambient temperature sampling point, the device is placed in a temperature chamber corresponding to that ambient temperature, and all temperature data from power-on to thermal stabilization are recorded. Each data point includes the following information (the temperature rise rate is defined as the temperature rise of the corresponding sensor within 5 minutes): current ambient temperature T. amb Current focal plane temperature T of the equipment jiaoCurrent device shutter temperature T s Current internal temperature T of the equipment j Current focal plane temperature rise rate S jiao Current device shutter temperature rise rate S s Current equipment cavity temperature rise rate S j Each data point constitutes a sample, and these samples together form a training dataset containing multiple samples.

[0043] In step 3, both the infrared image and the infrared thermal imager temperature data are acquired in real time, and the acquisition time for both is essentially the same. 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 temperature sensors inside the infrared thermal imager. These temperature sensors include, but are not limited to, a focal plane temperature sensor, a shutter temperature sensor, and an intracavity temperature sensor. The corresponding infrared device temperature data selects some temperature parameters related to 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 intracavity temperature, the device focal plane temperature rise rate, the device shutter temperature rise rate, and the device intracavity temperature rise rate. The device focal plane temperature rise rate, the device shutter temperature rise rate, and the device intracavity 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, or other related parameters may be included. To ensure prediction accuracy, the device temperature data must include at least two parameters.

[0044] The previous ambient temperature predictions were obtained through an open-loop prediction model. When faced with significant changes in external ambient temperature and the internal temperature sensors failing to update in time, the predictions will exhibit substantial deviations. Therefore, further verification of the prediction accuracy is necessary. Since the background temperature in the infrared image of the measured object should be largely consistent with the predicted ambient temperature, the accuracy of the infrared image of the measured object can be verified by checking their consistency.

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

[0046] The ambient temperature prediction module is only activated when the equipment temperature data is not fully stable. It uses an ambient temperature prediction model to obtain the current predicted ambient temperature value. Once the equipment temperature data is fully stable, the predicted ambient temperature value output by the model remains essentially unchanged. At this point, the ambient temperature prediction module can be deactivated, and the last predicted ambient temperature value obtained by the model can be used as the current predicted ambient temperature value, reducing equipment runtime and memory usage. The stability of a certain parameter in the equipment temperature data is defined as the range of the corresponding temperature sensor being less than a fixed threshold θ within a certain time period. For example, for the current shutter temperature, this means the shutter temperature transfer value within 5 minutes. When the ambient temperature prediction module is activated, θ is set to 0.5℃ in this embodiment.

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

[0048] The specific process of step 6 is further defined below:

[0049] Obtain the background radiation energy of the background region and the foreground radiation energy of the foreground region;

[0050] Based on the blackbody radiation law, the predicted ambient temperature is converted into reflected radiation.

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

[0052] The foreground radiation energy of the foreground region and the replaced background radiation energy are fed 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 regions to obtain the background region. There are many segmentation methods, such as thresholding and edge detection, which will not be elaborated upon here. Then, the background radiation energy is obtained from the segmented background region. Specifically, a radiation histogram is first calculated for the segmented background region.

[0054]

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

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

[0057] The radiation-temperature mapping curve is pre-calibrated and stored in the device. Using this radiation-temperature mapping curve, the background emitted radiation Y can be directly obtained. b ' g Obtain the corresponding infrared image background temperature T bg .

[0058] Of course, background radiation energy Y bg The effect is not only influenced by ambient temperature, but also by other factors such as the distance between the object being measured and the temperature measuring device. In this embodiment, the effect of ambient temperature on background radiation energy Y is only considered. bg Corrections are made to obtain the true background emitted radiation Y. b ' g In other embodiments, other factors may be considered for modification, and the modification methods are diverse and not limited to the methods provided in this embodiment.

[0059] The following further specifies the detailed process for obtaining background radiation energy:

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

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

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

[0063] The following describes the optimal method for the model: the model is implemented using a multiple linear regression model.

[0064] A multiple linear regression model is established, taking the temperature data of each component on the infrared thermal imager as input and the ambient temperature prediction value as output. The linear model is as follows:

[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 to β7 are model parameters;

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

[0068]

[0069] The loss function is defined as follows:

[0070]

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

[0072] The loss function is converged by gradient descent, and the model parameters β1 to β7 are finally obtained, which is the trained multiple linear regression model, and it is 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 downsampled data. Each downsampled data is projected into a space. The projections of each downsampled data onto multiple vectors in the space are multiplied to obtain each product result. Product results greater than a preset value are deleted. The remaining downsampled data corresponding to each product result are combined to form 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 shuffle the data according to the format required for training.

[0076] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A temperature correction method for an infrared thermal imager, characterized in that, The method includes the following steps: Step 1: Select n sampling temperatures uniformly from the set ambient temperature range, place the infrared thermal imager in the temperature chamber, set the temperature in the temperature chamber to n sampling temperatures in sequence, obtain the actual temperature data of each component on the infrared thermal imager at each set sampling temperature, process the actual temperature data of each component, and obtain the processed actual temperature data of each component. The actual temperature data of each component after processing and the corresponding set sampling temperature are used as a sample; the temperature data of each component after processing in the sample are used as the input of the ambient temperature prediction model, and the set sampling temperature is used as the output of the ambient temperature prediction model. Step 2: Train the ambient temperature prediction model using samples to obtain the trained ambient temperature prediction model; Step 3: Obtain the infrared image of the object being measured and the temperature data of each component on the infrared thermal imager. Input the temperature data of each component on the infrared thermal imager into the trained ambient temperature prediction model to obtain the predicted ambient temperature value. Step 4: Acquire the infrared image of the object being measured, wherein the infrared image of the object being measured includes a foreground region and a background region; Step 5: Determine whether the absolute value of the difference between the background area temperature and the predicted ambient temperature in the infrared image of the current object being measured is less than the preset value. If yes, output the infrared image temperature of the current object being measured as the final result. If no, proceed to step 6. Step 6: Replace the background area temperature in the infrared image temperature of the current object under test with the predicted ambient temperature value, 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 for an infrared thermal imager according to claim 1, characterized in that, In step 1, the temperature data of each component on the infrared thermal imager include: the focal plane temperature, the shutter temperature, the cavity temperature, the focal plane temperature rise rate, the shutter temperature rise rate, and the cavity temperature rise rate.

3. The temperature correction method for an infrared thermal imager according to claim 1, characterized in that, Step 6 is as follows: Obtain the background radiation energy of the background region and the foreground radiation energy of the foreground region; Based on the blackbody radiation law, the predicted ambient temperature is converted into reflected radiation. Replace the reflected radiation with the background radiation energy; The foreground radiation energy of the foreground region and the replaced background radiation energy are fed 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 for an infrared thermal imager according to claim 3, characterized in that, The specific process of obtaining the background radiation energy of the background region: Perform radiation histogram statistics on the background area; Convolve the radiation histogram to obtain the range of background radiation with the highest confidence level; The mean value of the background radiation range with the highest confidence level is taken as the background radiation energy.

5. The temperature correction method for an infrared thermal imager according to claim 1, characterized in that, Ambient temperature prediction is achieved using a multiple linear regression model.

6. The temperature correction method for an 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 downsampled data. Each downsampled data is projected into a space. The projections of each downsampled data onto multiple vectors in the space are multiplied to obtain each product result. Each product result is weighted and the weighted data is combined to form the temperature data of each component.

Images