An infrared thermal imager self-adaptive temperature drift compensation method and system

Abstract

The application provides an infrared thermal imager self-adaptive temperature drift compensation method and system, and relates to the technical field of infrared thermal imagers.The self-adaptive temperature drift compensation method comprises the following steps: collecting the real-time working temperature of a detector array based on a first temperature sensor built in an infrared thermal imager, and collecting the environmental temperature based on a second temperature sensor built in the infrared thermal imager; collecting the original infrared image output by the detector array, and calculating to obtain a target preliminary temperature; determining the current application scene based on the environmental temperature, the target preliminary temperature and the gray scale distribution characteristics of the infrared image; adaptively calling the corresponding temperature drift compensation model and parameters, calculating the gray scale correction amount of each pixel, correcting the original gray scale value based on the gray scale correction amount, and obtaining the compensated gray scale image; and calculating and outputting the target actual temperature based on the compensated gray scale image.The scheme effectively improves the scene adaptability of the infrared thermal imager, and solves the measurement deviation problem existing in different application scenes.

Description

Technical Field

[0001] This invention relates to the field of infrared thermal imager technology, and in particular to an adaptive temperature drift compensation method and system for infrared thermal imagers. Background Technology

[0002] Infrared thermal imagers, as advanced non-contact devices for measuring the surface temperature distribution of objects, play a crucial role in numerous fields such as industry, security, aerospace, and transportation. In industry, they are used for equipment fault diagnosis, quality inspection, and process monitoring, such as detecting overheating faults in electrical equipment in power systems and monitoring the temperature distribution of high-temperature furnaces in the metallurgical industry. In the security field, they can monitor and identify targets at night or in low light conditions, ensuring the safety of locations such as airports and borders. In aerospace and transportation, airborne or vehicle-mounted infrared thermal imagers can be used for navigation and target detection.

[0003] However, temperature drift is a prominent issue in practical applications. Temperature drift is a phenomenon where measurement results deviate due to changes in ambient temperature and the operating temperature of the detector array. Regarding ambient temperature, infrared thermal imagers often operate under varying temperature conditions, and large fluctuations affect detector performance. In cold outdoor environments or high-temperature industrial settings, temperature changes alter the detector's sensitivity and response characteristics, leading to deviations between measured grayscale values ​​and actual values, thus affecting accuracy. Furthermore, significant temperature differences across regions and seasons make it difficult for existing equipment to maintain stable performance. The operating temperature of the detector array itself is also critical. During operation, it heats up due to receiving infrared radiation, altering the physical properties of the materials, such as changes in semiconductor carrier concentration and mobility, affecting electrical performance and causing a shift in the relationship between the output electrical signal and the actual infrared radiation intensity, resulting in inaccurate results.

[0004] Existing temperature drift compensation methods are mostly simplistic and lack adaptability to complex scenarios. Some methods only consider one factor—the environment or the detector's operating temperature—resulting in limited compensation effectiveness. For example, methods based solely on ambient temperature compensation, ignoring the influence of the detector's own temperature variations, cannot effectively eliminate temperature drift during prolonged operation or under heavy loads. Different scenarios present vastly different operating conditions and requirements. Industrial temperature measurement demands high-precision measurement of high temperatures in relatively stable environments; outdoor security requires measuring low-temperature and normal-temperature targets in complex environments while dealing with interference; airborne or vehicle-mounted systems are affected by mechanical factors and experience drastic temperature changes. Existing methods are not optimized for these characteristics, making it difficult to achieve ideal compensation results and failing to meet the demands for high-precision, highly adaptable measurements.

[0005] In summary, existing infrared thermal imager temperature drift compensation technologies have significant shortcomings, failing to effectively address measurement deviation issues and lacking adaptability to various scenarios. Therefore, there is an urgent need to provide a solution to address these problems. Summary of the Invention

[0006] The purpose of this invention is to provide an adaptive temperature drift compensation method and system for infrared thermal imagers, so as to solve the problems that existing infrared thermal imagers cannot effectively solve the measurement deviation problem and lack scene adaptability.

[0007] The present invention provides an adaptive temperature drift compensation method for infrared thermal imagers, which adopts the following technical solution:

[0008] The real-time operating temperature of the detector array is collected based on the first temperature sensor built into the infrared thermal imager, and the ambient temperature is collected based on the second temperature sensor built into the infrared thermal imager.

[0009] The raw infrared image output by the detector array is acquired, the gray value of each pixel in the raw infrared image is extracted, and the initial temperature of the target is calculated based on the Stefan-Boltzmann law.

[0010] Based on preset scene type recognition rules, the current application scene is determined according to the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image;

[0011] Based on the identified application scenario type, the corresponding temperature drift compensation model and parameters are adaptively invoked. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration.

[0012] The real-time operating temperature, ambient temperature, grayscale value, and target preliminary temperature are input into the temperature drift compensation model that is adaptively called. The grayscale correction amount of each pixel is calculated, and the original grayscale value is corrected based on the grayscale correction amount to obtain the compensated grayscale image.

[0013] Based on the compensated grayscale image, combined with the corrected radiation energy temperature conversion formula, the actual temperature of the target is calculated and output.

[0014] Preferably, the multi-temperature point calibration process includes:

[0015] The infrared thermal imager was placed in a high and low temperature chamber with a temperature range of -40℃ to 85℃. Calibration temperature points were set at 10℃ intervals. The imager was aligned with a standard blackbody furnace, and the deviation data between the gray value at each calibration temperature point and the standard blackbody temperature was collected pixel by pixel. Based on the deviation data, a three-dimensional compensation function for each pixel was obtained.

[0016] Preferably, the formula for the three-dimensional compensation function is:

[0017] ;

[0018] in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, For the first The calibration coefficients of each pixel are obtained by solving the least squares method.

[0019] Preferably, the scene types include industrial temperature measurement scenes, outdoor security scenes, and airborne or vehicle-mounted scenes, and the scene type identification rules are as follows:

[0020] When the initial target temperature is greater than or equal to 100℃ and the ambient temperature fluctuation is less than or equal to 5℃ / min, it is determined to be an industrial temperature measurement scenario; when the initial target temperature is less than or equal to 50℃ and the grayscale variance of the infrared image is greater than or equal to 100, it is determined to be an outdoor security scenario; when the vibration acceleration detected by the IMU sensor built into the infrared thermal imager is greater than or equal to 0.5g, it is determined to be an airborne or vehicle-mounted scenario.

[0021] Preferably, the compensation model parameters for the adaptive invocation include:

[0022] When identified as an industrial temperature measurement scenario, a 4th-order polynomial optimized 3D compensation function is used, and the calibration temperature points are densified to one every 5°C. When identified as an outdoor security scenario, a simplified version of the 3D compensation function is used, and the noise suppression parameters are strengthened. When identified as an airborne or vehicle-mounted scenario, an anti-vibration compensation sub-model that integrates IMU data is called, and a vibration correction term is added.

[0023] Preferably, the formula for the fourth-order polynomial optimized three-dimensional compensation function is:

[0024] ;

[0025] The simplified formula for the three-dimensional compensation function is as follows:

[0026] ;

[0027] The compensation formula for the vibration-resistant compensation sub-model is as follows:

[0028]

[0029] in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, This is the vibration correction factor. Vibration acceleration detected by the IMU For the first Each pixel has its own calibration coefficient.

[0030] Preferably, before obtaining the compensated grayscale image, the process further includes:

[0031] Every 5-10 seconds, when the initial temperature of the target is detected to be no higher than the ambient temperature fluctuation threshold, a baseline image is acquired, the grayscale difference between the current frame and the baseline image is calculated, and the grayscale correction amount is adjusted based on the grayscale difference.

[0032] Preferably, after obtaining the compensated grayscale image, the method further includes:

[0033] The lens barrel temperature is collected by a third temperature sensor. When the difference between the lens barrel temperature and the reference temperature is greater than or equal to 10°C, the optical focal length compensation parameters are called to perform sharpening processing on the corrected grayscale image. The sharpening processing uses the Laplacian operator, and the operator coefficients are dynamically adjusted according to the difference in lens barrel temperature.

[0034] The preferred, modified radiative energy temperature conversion formula is:

[0035] ;

[0036] ;

[0037] in, The target actual temperature, The grayscale value after compensation. The detector response coefficient is dependent on ambient temperature. This represents the amount of dark signal drift caused by detector temperature. , , This is the response coefficient.

[0038] A second aspect of the present invention also provides an adaptive temperature drift compensation system for an infrared thermal imager, comprising a temperature acquisition module, a data acquisition module, a scene recognition module, a compensation model selection module, a temperature drift compensation module, and a data output module, wherein:

[0039] The temperature acquisition module is used to acquire the real-time operating temperature of the detector array based on the first temperature sensor built into the infrared thermal imager, and to acquire the ambient temperature based on the second temperature sensor built into the infrared thermal imager.

[0040] The data acquisition module is used to acquire the raw infrared image output by the detector array, extract the gray value of each pixel in the raw infrared image, and calculate the initial temperature of the target based on the Stefan-Boltzmann law.

[0041] The scene recognition module is used to determine the current application scene based on preset scene type recognition rules, the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image.

[0042] The compensation model selection module is used to adaptively call the corresponding temperature drift compensation model and parameters based on the identified application scenario type. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration.

[0043] The temperature drift compensation module is used to input the real-time working temperature, ambient temperature, gray value and target preliminary temperature into the temperature drift compensation model that is adaptively called, calculate the gray value correction amount for each pixel, correct the original gray value based on the gray value correction amount, and obtain the compensated gray value image.

[0044] The data output module is used to calculate and output the actual temperature of the target based on the compensated grayscale image and the corrected radiation energy temperature conversion formula.

[0045] The present invention provides an adaptive temperature drift compensation method and system for infrared thermal imagers, the advantages of which are as follows:

[0046] 1. This invention adopts a pixel-level three-dimensional compensation model, establishes a compensation function for each pixel through multi-temperature point calibration, and performs grayscale correction by comprehensively considering ambient temperature, detector operating temperature and target initial temperature. This solves the measurement deviation caused by the neglect of pixel-level temperature drift characteristics in existing methods due to global compensation or simplified models, and achieves higher precision temperature compensation.

[0047] 2. This invention optimizes the compensation model parameters for different scenarios. For example, a high-order polynomial model is used in industrial temperature measurement scenarios to improve accuracy, a simplified model is used in outdoor security scenarios to enhance noise suppression, and an anti-vibration compensation sub-model is introduced in airborne or vehicle-mounted scenarios. The corresponding temperature drift compensation model and parameters are adaptively called, which solves the measurement deviation problem caused by the fixed compensation model of existing infrared thermal imagers being unable to adapt to changing scenarios, and improves the accuracy of temperature measurement and system adaptability in different environments.

[0048] 3. This invention combines baseline image correction and optical focal length compensation technology. By acquiring baseline images at intervals to correct grayscale differences and dynamically adjusting sharpening processing according to lens temperature, it solves the additional deviations caused by sudden environmental changes or temperature drift of optical components in existing methods, and further enhances the robustness of temperature drift compensation and overall measurement consistency. Attached Figure Description

[0049] Figure 1 A flowchart provided for an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, 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. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art. The terms "comprising" and similar expressions used herein mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but does not exclude other elements or objects.

[0051] This invention provides an adaptive temperature drift compensation method for infrared thermal imagers, see [link to relevant documentation]. Figure 1 ,include:

[0052] S1. The real-time operating temperature of the detector array is collected based on the first temperature sensor built into the infrared thermal imager, and the ambient temperature is collected based on the second temperature sensor built into the infrared thermal imager.

[0053] S2. Acquire the raw infrared image output by the detector array, extract the grayscale value of each pixel in the raw infrared image, and calculate the initial temperature of the target based on the Stefan-Boltzmann law;

[0054] S3. Based on preset scene type recognition rules, determine the current application scene according to the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image;

[0055] S4. Based on the identified application scenario type, the corresponding temperature drift compensation model and parameters are adaptively called. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration.

[0056] S5. Input the real-time working temperature, ambient temperature, gray value and target preliminary temperature into the temperature drift compensation model that is adaptively called, calculate the gray value correction amount for each pixel, correct the original gray value based on the gray value correction amount, and obtain the compensated gray value image.

[0057] S6. Based on the compensated grayscale image and the corrected radiation energy temperature conversion formula, calculate and output the actual temperature of the target.

[0058] In some embodiments, the process of performing step S1 includes: the first temperature sensor is a PT1000 high-precision temperature sensor, which can be directly attached to the substrate surface of the detector array or encapsulated inside the detector cavity, forming a tight thermal conduction contact with the core sensitive area of ​​the detector array, ensuring that the real temperature state of the detector array during operation can be captured in real time, so as to avoid temperature conduction delay or deviation caused by improper installation position.

[0059] Furthermore, the second temperature sensor is an NTC thermistor, which is installed in the area inside the thermal imager housing near the optical components and is thermally isolated from the constant temperature cavity of the detector array. This not only avoids interference from the heat generated by the detector array during operation on the acquisition of ambient temperature, but also allows for rapid response to changes in the external ambient temperature. At the same time, the sensor is wrapped in a protective shell to prevent dust, moisture, and other factors from affecting the sensor's performance and to ensure the stability of acquisition in harsh environments.

[0060] In some embodiments, the process of performing step S2 includes: receiving infrared radiation from the target based on the detector array, converting it into an electrical signal, amplifying, filtering, and performing AD conversion to generate an original infrared digital image, reading the digital signal of each pixel in the image in row and column order, and converting it into a grayscale value using a calibration mapping table. Remove bad pixels and fill them with interpolation.

[0061] Furthermore, the target region is extracted through image segmentation, and the average grayscale value of pixels within the region is calculated. Substituting the simplified version of the Stefan-Boltzmann law into the formula, and solving for the fourth root, we obtain the target initial temperature. The simplified version of the Stefan-Boltzmann law is as follows:

[0062] ;

[0063] in, The initial target temperature, Grayscale value To calculate the average grayscale value of pixels within the region, For ambient temperature, The initial response coefficient of the detector is given at a reference ambient temperature of 25°C.

[0064] Furthermore, Through multi-temperature calibration experiments before shipment, it was determined that, with an ambient temperature of 25℃ in the high and low temperature chamber and using the known radiation energy output of a standard blackbody furnace as a benchmark, the grayscale value output of the detector array was collected. The formula "Radiation energy = Grayscale value / The correspondence between "" and "" was fitted to obtain A fixed value.

[0065] In practice, since the initial target temperature is only used for subsequent scene recognition, a fixed temperature is used. This simplifies the calculation process, improves computational efficiency, and avoids real-time performance losses caused by calling complex dynamic coefficients; subsequently, high-precision temperature calculations are pursued through follow-up steps. Replace with dynamically adjusted based on ambient temperature. This results in a hierarchical design with "fixed baseline coefficients for preliminary calculations and dynamic adaptation coefficients for precise calculations," which balances efficiency and accuracy, and conforms to the logic of overall adaptive compensation.

[0066] In some embodiments, the process of performing step S3 includes: determining the current application scenario based on preset scene type identification rules, wherein the scene type includes industrial temperature measurement scenario, outdoor security scenario and airborne or vehicle-mounted scenario, and based on the ambient temperature, the initial temperature of the target and the grayscale distribution characteristics of the infrared image.

[0067] Furthermore, the scene type identification rules are as follows: when the initial target temperature is greater than or equal to 100℃ and the ambient temperature fluctuation is less than or equal to 5℃ / min, it is determined to be an industrial temperature measurement scene; when the initial target temperature is less than or equal to 50℃ and the infrared image grayscale variance is greater than or equal to 100, it is determined to be an outdoor security scene; when the vibration acceleration detected by the IMU sensor built into the infrared thermal imager is greater than or equal to 0.5g, it is determined to be an airborne or vehicle-mounted scene.

[0068] In some embodiments, the process of executing step S4 includes: using a pixel-level three-dimensional compensation model established through multi-temperature point calibration, and adaptively calling the corresponding three-dimensional temperature drift compensation model and parameters based on the identified application scenario type.

[0069] Furthermore, the process of establishing a pixel-level three-dimensional compensation model through multi-temperature point calibration includes: placing an infrared thermal imager in a high-low temperature chamber with a temperature range of -40℃ to 85℃, setting calibration temperature points at 10℃ intervals, aligning it with a standard blackbody furnace, collecting pixel-by-pixel deviation data between the grayscale value at each calibration temperature point and the standard blackbody temperature, and fitting a three-dimensional compensation function for each pixel based on the deviation data. The formula for the three-dimensional compensation function is:

[0070] ;

[0071] in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, For the first The calibration coefficients of each pixel are obtained by solving the least squares method.

[0072] Furthermore, based on the identified application scenario type, the process of adaptively calling the corresponding three-dimensional temperature drift compensation model and parameters includes: when identified as an industrial temperature measurement scenario, a four-order polynomial optimized three-dimensional compensation function is used, and the calibration temperature points are densified to one every 5°C; when identified as an outdoor security scenario, a simplified version of the three-dimensional compensation function is used, and noise suppression parameters are strengthened; when identified as an airborne or vehicle-mounted scenario, an anti-vibration compensation sub-model that integrates IMU data is called, and vibration correction terms are added.

[0073] In practice, by comprehensively analyzing ambient temperature, target initial temperature, and image grayscale features, the system automatically identifies typical scenarios such as industrial temperature measurement, outdoor security, and airborne / vehicle-mounted applications, and seamlessly switches to the corresponding optimal compensation model. This allows for the use of more complex fourth-order polynomial models in high-precision industrial applications, simplified and efficient models in real-time security applications, and vibration-resistant sub-models fused with IMU data in vibration environments. This completely overcomes the limitations of traditional single compensation models, ensuring optimal temperature measurement performance under each specific operating condition.

[0074] Furthermore, the formula for the three-dimensional compensation function optimized by the fourth-order polynomial is as follows:

[0075] ;

[0076] The simplified formula for the three-dimensional compensation function is as follows:

[0077] ;

[0078] The compensation formula for the vibration-resistant compensation sub-model is as follows:

[0079]

[0080] in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, This is the vibration correction factor. Vibration acceleration detected by the IMU For the first Each pixel has its own calibration coefficient.

[0081] In some embodiments, the process of executing step S5 includes: inputting the real-time operating temperature, ambient temperature, grayscale value and target preliminary temperature into the temperature drift compensation model that is adaptively invoked, calculating the grayscale correction amount for each pixel, and correcting the original grayscale value based on the grayscale correction amount.

[0082] Furthermore, every 5 to 10 seconds when the initial temperature of the target is detected to be no higher than the ambient temperature fluctuation threshold, a baseline image is acquired, the grayscale difference between the current frame and the baseline image is calculated, and the grayscale correction amount is corrected based on the grayscale difference to obtain the compensated grayscale image.

[0083] In fact, after obtaining the compensated grayscale image, the process also includes: collecting the lens barrel temperature through a third temperature sensor; when the difference between the lens barrel temperature and the reference temperature is greater than or equal to 10°C, calling the optical focal length compensation parameters to perform sharpening processing on the corrected grayscale image; the sharpening processing uses the Laplacian operator, and the operator coefficients are dynamically adjusted according to the difference in the lens barrel temperature.

[0084] In fact, periodic baseline image comparison can effectively suppress the slow drift of ambient background radiation, while image sharpening processing based on dynamic adjustment of the lens barrel temperature can compensate for image quality degradation caused by the thermal effects of the optical system in the digital domain. These measures, together with the core temperature drift compensation model, jointly address various interference factors that affect measurement results, enabling the infrared thermal imager to maintain stable high-performance output even in more demanding and variable working environments.

[0085] In some embodiments, the process of performing step S6 includes: calculating and outputting the actual target temperature based on the compensated grayscale image and in conjunction with the corrected radiation energy temperature conversion formula.

[0086] Furthermore, the revised formula for converting radiative energy to temperature is as follows:

[0087] ;

[0088] ;

[0089] in, The target actual temperature, The grayscale value after compensation. The detector response coefficient is dependent on ambient temperature. This represents the amount of dark signal drift caused by detector temperature. , , These are the response coefficients obtained through calibration and fitting at multiple ambient temperature points.

[0090] In fact, this invention employs a simplified formula with fixed coefficients in the initial scene recognition and target preliminary temperature calculation stages to quickly complete scene judgment, saving valuable time for subsequent decision-making. In the final temperature inversion stage, it calls upon a precise coefficient model that dynamically changes with ambient temperature for high-precision calculations. This design ensures both the system's rapid response to changes in the external environment and the extremely high accuracy of the final output, making this method suitable for dynamic applications with stringent requirements for both real-time performance and accuracy.

[0091] A second aspect of the present invention also provides an adaptive temperature drift compensation system for an infrared thermal imager, comprising a temperature acquisition module, a data acquisition module, a scene recognition module, a compensation model selection module, a temperature drift compensation module, and a data output module, wherein:

[0092] The temperature acquisition module is used to acquire the real-time operating temperature of the detector array based on the first temperature sensor built into the infrared thermal imager, and to acquire the ambient temperature based on the second temperature sensor built into the infrared thermal imager.

[0093] The data acquisition module is used to acquire the raw infrared image output by the detector array, extract the gray value of each pixel in the raw infrared image, and calculate the initial temperature of the target based on the Stefan-Boltzmann law.

[0094] The scene recognition module is used to determine the current application scene based on preset scene type recognition rules, the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image.

[0095] The compensation model selection module is used to adaptively call the corresponding temperature drift compensation model and parameters based on the identified application scenario type. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration.

[0096] The temperature drift compensation module is used to input the real-time working temperature, ambient temperature, gray value and target preliminary temperature into the temperature drift compensation model that is adaptively called, calculate the gray value correction amount for each pixel, correct the original gray value based on the gray value correction amount, and obtain the compensated gray value image.

[0097] The data output module is used to calculate and output the actual temperature of the target based on the compensated grayscale image and the corrected radiation energy temperature conversion formula.

[0098] While embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations fall within the scope and spirit of the invention as set forth in the claims. Furthermore, the invention described herein may have other embodiments and can be implemented or carried out in various ways.

Claims

1. An adaptive temperature drift compensation method for infrared thermal imagers, characterized in that, Includes the following steps: The real-time operating temperature of the detector array is collected based on the first temperature sensor built into the infrared thermal imager, and the ambient temperature is collected based on the second temperature sensor built into the infrared thermal imager. The raw infrared image output by the detector array is acquired, the gray value of each pixel in the raw infrared image is extracted, and the initial temperature of the target is calculated based on the Stefan-Boltzmann law. Based on preset scene type recognition rules, the current application scene is determined according to the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image; Based on the identified application scenario type, the corresponding temperature drift compensation model and parameters are adaptively invoked. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration. The real-time operating temperature, ambient temperature, grayscale value, and target preliminary temperature are input into the temperature drift compensation model that is adaptively called. The grayscale correction amount of each pixel is calculated, and the original grayscale value is corrected based on the grayscale correction amount to obtain the compensated grayscale image. Based on the compensated grayscale image, combined with the corrected radiation energy temperature conversion formula, the actual temperature of the target is calculated and output.

2. The adaptive temperature drift compensation method for infrared thermal imagers according to claim 1, characterized in that, The multi-temperature point calibration process includes: The infrared thermal imager was placed in a high and low temperature chamber with a temperature range of -40℃ to 85℃. Calibration temperature points were set at 10℃ intervals. The imager was aligned with a standard blackbody furnace, and the deviation data between the gray value at each calibration temperature point and the standard blackbody temperature was collected pixel by pixel. Based on the deviation data, a three-dimensional compensation function for each pixel was obtained.

3. The adaptive temperature drift compensation method for infrared thermal imagers according to claim 2, characterized in that, The formula for the three-dimensional compensation function is: ； in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, For the first The calibration coefficients of each pixel are obtained by solving the least squares method.

4. The adaptive temperature drift compensation method for infrared thermal imagers according to claim 1, characterized in that, The scenario types include industrial temperature measurement scenarios, outdoor security scenarios, and airborne or vehicle-mounted scenarios. The scenario type identification rules are as follows: When the initial target temperature is greater than or equal to 100℃ and the ambient temperature fluctuation is less than or equal to 5℃ / min, it is determined to be an industrial temperature measurement scenario. When the initial target temperature is less than or equal to 50°C and the grayscale variance of the infrared image is greater than or equal to 100, it is determined to be an outdoor security scene; When the vibration acceleration detected by the IMU sensor built into the infrared thermal imager is greater than or equal to 0.5g, it is determined to be an airborne or vehicle-mounted scenario.

5. The adaptive temperature drift compensation method for infrared thermal imagers according to claim 4, characterized in that, The parameters of the adaptive invocation compensation model include: When the scenario is identified as an industrial temperature measurement scenario, a three-dimensional compensation function optimized by a fourth-order polynomial is used, and the calibration temperature points are densified to one every 5°C. When the scene is identified as an outdoor security scenario, a simplified version of the 3D compensation function is used to enhance the noise suppression parameters. When the scenario is identified as airborne or vehicle-mounted, the vibration compensation sub-model that integrates IMU data is invoked to add a vibration correction term.

6. The adaptive temperature drift compensation method for an infrared thermal imager according to claim 5, characterized in that, include: The formula for the three-dimensional compensation function optimized by the fourth-order polynomial is as follows: ； The simplified formula for the three-dimensional compensation function is as follows: ； The compensation formula for the vibration-resistant compensation sub-model is as follows: ； in, For the first Grayscale compensation correction amount per pixel For ambient temperature, For real-time operating temperature, The initial target temperature, This is the vibration correction factor. Vibration acceleration detected by the IMU For the first Each pixel has its own calibration coefficient.

7. The adaptive temperature drift compensation method for infrared thermal imagers according to claim 1, characterized in that, Before obtaining the compensated grayscale image, the process also includes: Every 5-10 seconds, when the initial temperature of the target is detected to be no higher than the ambient temperature fluctuation threshold, a baseline image is acquired, the grayscale difference between the current frame and the baseline image is calculated, and the grayscale correction amount is adjusted based on the grayscale difference.

8. The adaptive temperature drift compensation method for an infrared thermal imager according to claim 1, characterized in that, After obtaining the compensated grayscale image, it also includes: The lens barrel temperature is collected by a third temperature sensor. When the difference between the lens barrel temperature and the reference temperature is greater than or equal to 10°C, the optical focal length compensation parameters are called to perform sharpening processing on the corrected grayscale image. The sharpening processing uses the Laplacian operator, and the operator coefficients are dynamically adjusted according to the difference in lens barrel temperature.

9. The adaptive temperature drift compensation method for an infrared thermal imager according to claim 1, characterized in that, The revised formula for converting radiative energy to temperature is: ； ； in, The target actual temperature, The grayscale value after compensation. The detector response coefficient is dependent on ambient temperature. This represents the amount of dark signal drift caused by detector temperature. , , This is the response coefficient.

10. An adaptive temperature drift compensation system for an infrared thermal imager, characterized in that, It includes a temperature acquisition module, a data acquisition module, a scene recognition module, a compensation model selection module, a temperature drift compensation module, and a data output module, among which: The temperature acquisition module is used to acquire the real-time operating temperature of the detector array based on the first temperature sensor built into the infrared thermal imager, and to acquire the ambient temperature based on the second temperature sensor built into the infrared thermal imager. The data acquisition module is used to acquire the raw infrared image output by the detector array, extract the gray value of each pixel in the raw infrared image, and calculate the initial temperature of the target based on the Stefan-Boltzmann law. The scene recognition module is used to determine the current application scene based on preset scene type recognition rules, the ambient temperature, the initial temperature of the target, and the grayscale distribution characteristics of the infrared image. The compensation model selection module is used to adaptively call the corresponding temperature drift compensation model and parameters based on the identified application scenario type. The temperature drift compensation model is a pixel-level three-dimensional compensation model that has been pre-established through multi-temperature point calibration. The temperature drift compensation module is used to input the real-time working temperature, ambient temperature, gray value and target preliminary temperature into the temperature drift compensation model that is adaptively called, calculate the gray value correction amount for each pixel, correct the original gray value based on the gray value correction amount, and obtain the compensated gray value image. The data output module is used to calculate and output the actual temperature of the target based on the compensated grayscale image and the corrected radiation energy temperature conversion formula.

Images