Intelligent calibration method and system for infrared thermal imager

Abstract

The application relates to the technical field of intelligent calibration, and discloses an intelligent calibration method and system for an infrared thermal imager, the method comprising the following steps: acquiring a temperature distribution image and acceleration data; performing posture angle calculation according to the acceleration data to obtain posture angle data; performing temperature difference calculation according to the temperature distribution image to obtain an initial temperature difference; performing neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight; performing abnormal point detection according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain abnormal point data; and performing calibration optimization according to the abnormal point data and the posture angle data to output a calibration temperature image. The method has the following effects: the calibration precision of the infrared thermal imager can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent calibration, in particular to an intelligent calibration method and system for an infrared thermal imager. BACKGROUND

[0002] With the development of technology, infrared thermal imaging technology has been widely used in many fields such as security monitoring, industrial detection and medical diagnosis. For example, in power maintenance, the temperature change of equipment can be monitored in real time by an infrared thermal imager, and overheating points can be found in time to prevent faults. However, due to changes in environmental temperature, precision drift of the instrument itself, and aging of the detector, the infrared thermal imager will have measurement errors, affecting its accuracy. Especially in complex working environments, such as high-altitude operations or places with limited space, the influence of these factors is more significant. Therefore, how to intelligently calibrate the infrared thermal imager to ensure its long-term stable operation and provide accurate temperature measurement results has become an important research direction in this field.

[0003] An existing intelligent calibration method for an infrared thermal imager is to combine an internal reference source with an external standard blackbody. First, an internal reference source with a known temperature is set in the instrument as a fixed reference point in the calibration process. Then, in actual use, an external standard blackbody is placed in the environment to be measured, and its temperature is kept constant. Then, the infrared thermal imager measures the temperature values of the internal reference source and the external blackbody at the same time, and automatically adjusts the gain and offset parameters of the instrument according to the difference between the two, thereby realizing calibration. This method can compensate for measurement errors caused by changes in environmental temperature by comparing internal and external temperature readings, improving the accuracy of temperature measurement.

[0004] Although the above method solves the calibration problem of the infrared thermal imager to some extent, there are still some deficiencies. One of the important problems is that the existing method does not take into account the influence of spatial relationships such as device attitude angle on measurement results. When the infrared thermal imager measures at different angles, the data obtained is biased due to different viewing angles, resulting in low calibration accuracy of the thermal imager. SUMMARY

[0005] The present application provides an intelligent calibration method and system for an infrared thermal imager to improve the calibration accuracy of the thermal imager.

[0006] In a first aspect, to solve the above technical problems, the present application provides an intelligent calibration method for an infrared thermal imager, comprising:

[0007] obtaining a temperature distribution image and acceleration data;

[0008] According to the acceleration data, a posture angle is calculated to obtain posture angle data;

[0009] According to the temperature distribution image, a temperature difference is calculated to obtain an initial temperature difference;

[0010] According to the temperature distribution image and the initial temperature difference, neighborhood analysis is performed to obtain an adaptive weight;

[0011] According to the adaptive weight, the initial temperature difference, and a preset standard deviation threshold, an abnormal point is detected to obtain abnormal point data;

[0012] According to the abnormal point data and the posture angle data, calibration optimization is performed to output a calibrated temperature image.

[0013] In an optional implementation, the calculating of the posture angle according to the acceleration data to obtain the posture angle data comprises:

[0014] The pitch angle is calculated by the following formula:

[0015]

[0016] The roll angle is calculated by the following formula:

[0017]

[0018] wherein, the pitch angle is denoted as the roll angle is denoted as the acceleration component of the axis, the acceleration component of the axis, the acceleration component of the axis.

[0019] The posture angle data comprises the pitch angle and the roll angle.

[0020] In an optional implementation, the calculating of the temperature difference according to the temperature distribution image to obtain the initial temperature difference comprises:

[0021] The horizontal temperature difference is calculated by the following formula:

[0022]

[0023] wherein, the horizontal temperature difference of a pixel is denoted as the temperature value of a pixel is denoted as the temperature value of a pixel is denoted as

[0024] The vertical temperature difference is calculated by the following formula:

[0025]

[0026] wherein, represents the vertical temperature difference of the pixel , represents the temperature value of the pixel , represents the temperature value of the pixel ;

[0027] The temperature change direction is calculated by the following formula:

[0028]

[0029] wherein, represents the temperature change direction of the pixel ;

[0030] The initial temperature difference comprises the temperature change direction, the horizontal temperature difference and the vertical temperature difference.

[0031] In an optional embodiment, the neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight comprises:

[0032] The temperature distribution image is boundary filled to obtain a filled temperature image;

[0033] The variance is calculated according to a preset neighborhood rule and the filled temperature image to obtain a neighborhood variance;

[0034] The temperature gradient is calculated according to the initial temperature difference to obtain a temperature gradient:

[0035] The weight is calculated by the following formula:

[0036]

[0037] wherein, represents the adaptive weight of the pixel , represents the temperature gradient of the pixel , represents the neighborhood variance.

[0038] In an optional embodiment, the abnormal point detection according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain abnormal point data comprises:

[0039] The temperature difference is updated according to the adaptive weight and the initial temperature difference to obtain an updated temperature difference;

[0040] Get the temperature difference magnitude threshold and temperature difference direction threshold;

[0041] When the updated horizontal temperature difference or the updated vertical temperature difference of the updated temperature difference is greater than the temperature difference threshold, the next step is determined; otherwise, an abnormal point detection is performed on the next pixel;

[0042] When the updated change direction of the updated temperature difference is greater than the temperature difference direction threshold, the next step is determined; otherwise, an abnormal point detection is performed on the next pixel;

[0043] The updated standard deviation is calculated using the following formula:

[0044]

[0045] in, Represents pixels The updated standard deviation of represents the number of neighborhood pixels, Represents pixels Neighborhood, Represents pixels The updated horizontal temperature difference, Represents pixels The updated vertical temperature difference, represents the average of the updated horizontal temperature differences within the neighborhood, represents the average of the updated vertical temperature differences within the neighborhood;

[0046] When the updated standard deviation is greater than the standard deviation threshold, the pixel is determined to be an abnormal point;

[0047] The outlier data includes all the outliers.

[0048] In an optional embodiment, performing calibration optimization based on the abnormal point data and the attitude angle data and outputting a calibrated temperature image includes:

[0049] Performing data filling on the abnormal point data to obtain an abnormality corrected image;

[0050] The calibration temperature of the calibration temperature image is calculated by the following formula:

[0051]

[0052] in, Represents pixels The calibration temperature, Indicates the pixels in the abnormally corrected image The corrected temperature value, represents the pitch angle, represents the roll angle;

[0053] generating a calibration temperature image according to the calibration temperature, and outputting the calibration temperature image.

[0054] In an optional implementation, the data padding on the abnormal point data to obtain an abnormal correction image comprises:

[0055] replacing the temperature value of the abnormal point data with a neighborhood temperature average value.

[0056] In a second aspect, the present application provides an intelligent calibration system of an infrared thermal imager, comprising:

[0057] a data acquisition module configured to acquire a temperature distribution image and acceleration data;

[0058] a posture analysis module configured to calculate a posture angle according to the acceleration data to obtain posture angle data;

[0059] a temperature difference calculation module configured to calculate a temperature difference according to the temperature distribution image to obtain an initial temperature difference;

[0060] a weight calculation module configured to perform neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight;

[0061] an abnormality detection module configured to detect abnormal points according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain abnormal point data;

[0062] a temperature calibration module configured to perform calibration optimization according to the abnormal point data and the posture angle data to output a calibration temperature image.

[0063] In a third aspect, the present application further provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the intelligent calibration method of the infrared thermal imager according to any one of the above descriptions when executing the computer program.

[0064] In a fourth aspect, the present application further provides a computer readable storage medium comprising a stored computer program, wherein the computer readable storage medium controls a device where the computer readable storage medium is located to execute the intelligent calibration method of the infrared thermal imager according to any one of the above descriptions when the computer program runs.

[0065] Compared with the prior art, the present application has the following beneficial effects: the present application discloses an intelligent calibration method and system for an infrared thermal imager, the method comprising acquiring a temperature distribution image and acceleration data; performing attitude angle calculation according to the acceleration data to obtain attitude angle data; performing temperature difference calculation according to the temperature distribution image to obtain an initial temperature difference; performing neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight; performing abnormal point detection according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain abnormal point data; and performing calibration optimization according to the abnormal point data and the attitude angle data to output a calibrated temperature image. The method has the following effects: the method can improve the calibration accuracy of the infrared thermal imager.

[0066] Specifically, the present method determines the angle of the device relative to the direction of gravity based on the three-dimensional acceleration values output by the accelerometer, which has significant technical advantages. First, this calculation method directly utilizes the data provided by the physical sensor, without the need for additional hardware support, simplifying system design and improving computational efficiency. Second, the use of two-dimensional formulas ensures the accuracy of angle calculation, providing reliable attitude estimation even in dynamic environments. The application of these formulas enables the infrared thermal imager to automatically adjust its viewing angle and perform image correction, thereby improving image quality and calibration accuracy. Furthermore, the above formulas are derived based on coordinate transformation principles and trigonometric relationships, consistent with the attitude description method in rigid body dynamics. Compared with traditional calibration methods, this method not only improves the automation level of the calibration process, but also enhances the adaptability to different environmental conditions.

[0067] Further, the present method performs neighborhood analysis and calculates adaptive weights based on the temperature distribution image and the initial temperature difference. This process includes boundary filling of the temperature distribution image to obtain a filled temperature image, followed by variance calculation based on the preset neighborhood rules and the filled temperature image, thereby obtaining neighborhood variance. Further, gradient calculation is performed based on the initial temperature difference, resulting in the temperature gradient of each pixel point. Finally, adaptive weights are calculated through the formula.

[0068] This adaptive weight calculation method has multiple advantages. First, by performing boundary filling on the temperature distribution image, edge effects can be effectively avoided, ensuring the data integrity of the edge region and making the overall temperature field analysis more accurate. Second, the introduction of neighborhood variance takes into account the temperature changes within the local region, helping to more accurately capture subtle temperature differences and improve calibration accuracy. In addition, the calculation of temperature gradient reflects the rate of temperature change with space, combined with the exponential function form of the weight calculation formula, it can suppress noise while preserving important features, enhancing the stability of the image processing results.

[0069] Further, the rationality of the formula is reflected in its use of an exponential decay form to define the weight, that is, the greater the modulus square of the temperature gradient, the more intense the temperature change, and the smaller the corresponding weight, and vice versa. This method can naturally emphasize areas with relatively gentle temperature changes while reducing the influence of areas with rapid changes or noise interference. Compared with traditional methods, this adaptive weighting strategy not only improves the adaptability of the infrared thermal imager to different scenes during calibration, but also significantly improves the quality and detection accuracy of the final image.

[0070] Further, the method proposes a step of detecting abnormal points by adaptive weight, initial temperature difference and preset standard deviation threshold, which can effectively improve the image quality and the accuracy of subsequent analysis. This step first updates the temperature difference based on the adaptive weight and the initial temperature difference to obtain an updated temperature difference. This step aims to adjust the temperature difference of each pixel point according to the environmental information around it to more accurately reflect the true situation. The temperature difference size and temperature difference change direction are first screened to reduce the subsequent calculation amount and improve the detection efficiency.

[0071] Further, when the updated standard deviation of a certain pixel point exceeds the preset standard deviation threshold, the pixel point is identified as an abnormal point and is included in the abnormal point data set. This method ensures the accuracy and robustness of abnormal point detection through a multi-level filtering mechanism (including comparison of temperature difference size, direction and standard deviation). Compared with traditional single threshold judgment methods, it is more suitable for complex actual scenes, reduces the false positive rate and false negative rate, and thus improves the performance and reliability of the entire infrared thermal imaging system.

[0072] Further, the abnormal point data is processed and combined with the pose angle data to optimize the calibration process. This method first fills the abnormal point data by replacing the temperature values of abnormal points with the average temperature values of the neighborhood to generate an abnormal correction image. This method effectively reduces the influence of abnormal values on the quality of the final temperature image and improves the accuracy and reliability of the temperature distribution. By considering the device's pose angle, the formula can compensate for measurement errors caused by device tilt, ensuring the accuracy of the measured temperature under different poses. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 is a flowchart of an intelligent calibration method of an infrared thermal imager provided by the first embodiment of the present application;

[0074] Figure 2 is a structural diagram of an intelligent calibration system of an infrared thermal imager provided by the second embodiment of the present application. DETAILED DESCRIPTION

[0075] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0076] With the development of technology, infrared thermal imaging technology has been widely applied in many fields such as security monitoring, industrial detection and medical diagnosis. For example, in power maintenance, the temperature change of equipment can be monitored in real time by an infrared thermal imager, and overheating points can be found in time to prevent faults. However, due to factors such as changes in environmental temperature, precision drift of the instrument itself, and aging of the detector, the infrared thermal imager will have measurement errors, affecting its accuracy. Especially in complex working environments, such as high-altitude operations or places with limited space, the influence of these factors is more significant. Therefore, how to intelligently calibrate the infrared thermal imager to ensure its long-term stable operation and provide accurate temperature measurement results has become an important research direction in this field.

[0077] An existing intelligent calibration method for infrared thermal imagers is to combine an internal reference source with an external standard blackbody. First, an internal reference source with a known temperature is set in the instrument as a fixed reference point in the calibration process. Then, in actual use, an external standard blackbody is placed in the environment to be measured, and its temperature is kept constant. Then, the infrared thermal imager measures the temperature values of the internal reference source and the external blackbody at the same time, and automatically adjusts the gain and offset parameters of the instrument according to the difference between the two, thereby realizing calibration. This method can compensate for measurement errors caused by changes in environmental temperature by comparing internal and external temperature readings, improving the accuracy of temperature measurement.

[0078] Although the above method solves the calibration problem of the infrared thermal imager to some extent, there are still some deficiencies. One of the important problems is that the existing method does not consider the influence of spatial relationships such as device attitude angle on measurement results. When the infrared thermal imager measures at different angles, the data obtained is biased due to different viewing angles, resulting in low calibration accuracy of the thermal imager.

[0079] Reference Figure 1 The first embodiment of the present application provides an intelligent calibration method for an infrared thermal imager, comprising the following steps:

[0080] S11, acquiring a temperature distribution image and acceleration data;

[0081] S12, performing attitude angle calculation according to the acceleration data to obtain attitude angle data;

[0082] S13, performing temperature difference calculation according to the temperature distribution image to obtain an initial temperature difference;

[0083] S14, performing neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight;

[0084] S15, performing abnormal point detection according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain abnormal point data;

[0085] S16, performing calibration optimization according to the abnormal point data and the attitude angle data to output a calibrated temperature image.

[0086] In step S11, a temperature distribution image and acceleration data are acquired.

[0087] In an embodiment, the temperature distribution image and the acceleration data are acquired by an infrared thermal imager and a built-in three-axis accelerometer. For the acquisition of the temperature distribution image, the basic parameters of the thermal imager need to be set first, for example, the frame rate can be set to 30 Hz to ensure that the temperature changes in a dynamic scene can be accurately captured, the resolution is set to 640x480 to ensure the image quality while taking into account the speed of data processing, and the temperature range is selected according to the actual application scenario, for example, set to -20℃ to +500℃. In addition, the focal length needs to be adjusted to ensure that the target area is clearly imaged, and the automatic gain control (AGC) function is turned on to optimize the visualization effect in different temperature ranges. For the acquisition of the acceleration data, the sampling frequency of the accelerometer needs to be configured to be synchronized with the frame rate of the thermal imager, i.e. 30 Hz, so as to ensure the data matching when calculating the attitude angle; the measurement range is set according to the maximum acceleration experienced by the device, such as ±16g, and the sensitivity needs to be adjusted according to specific requirements, which can be set to 1 mg / LSB. After the acquisition is completed, in order to ensure the accuracy of subsequent processing, the original temperature distribution image needs to be preprocessed, including noise removal, contrast enhancement and other operations, and the acceleration data needs to be checked for its integrity and consistency, and if necessary, filtering processing is performed to reduce the influence of high-frequency noise. The data obtained after such processing is used in the subsequent attitude angle calculation and temperature difference analysis processes.

[0088] In step S12, attitude angle calculation is performed according to the acceleration data to obtain attitude angle data.

[0089] In an embodiment, the pitch angle is calculated by the following formula:

[0090]

[0091] The roll angle is calculated by the following formula:

[0092]

[0093] wherein, the pitch angle, the roll angle, the acceleration component of the axis, the acceleration component of the axis, the acceleration component of the axis;

[0094] wherein the attitude angle data comprises the pitch angle and the roll angle.

[0095] It is worth noting that the pitch angle describes the degree of inclination of the device along the axis, i.e. the angle of forward and backward inclination. When the front of the device is raised upward, the pitch angle is positive; otherwise, when the front is downward, the pitch angle is negative. The pitch angle helps determine whether the device is looking upward or downward. The roll angle describes the degree of inclination of the device along the axis, i.e. the angle of left and right inclination. When the right side of the device is raised upward, the roll angle is positive; otherwise, when the left side is upward, the roll angle is negative. The roll angle is used to determine whether the device is horizontal and the direction of its inclination.

[0096] In step S13, the initial temperature difference is obtained by performing temperature difference calculation according to the temperature distribution image.

[0097] In one embodiment, the horizontal temperature difference is calculated by the following formula:

[0098]

[0099] wherein, the horizontal temperature difference of the pixel , T, the temperature value of the pixel , T, the temperature value of the pixel , T;

[0100] The vertical temperature difference is calculated by the following formula:

[0101]

[0102] wherein, the vertical temperature difference of the pixel , T, the temperature value of the pixel , T, the temperature value of the pixel , T;

[0103] The temperature change direction is calculated by the following formula:

[0104]

[0105] wherein, represents the temperature change direction of a pixel ;

[0106] wherein the initial temperature difference comprises the temperature change direction, the horizontal temperature difference and the vertical temperature difference.

[0107] It is worth noting that the horizontal temperature difference calculates the average temperature difference between each pixel point and its left and right adjacent pixel points, i.e. the temperature gradient in the horizontal direction. A smooth temperature change rate is obtained by dividing the temperature value of the right pixel point by the temperature value of the left pixel point by 2. The vertical temperature difference calculates the average temperature difference between each pixel point and its upper and lower adjacent pixel points, i.e. the temperature gradient in the vertical direction. A smooth temperature change rate is obtained by dividing the temperature value of the upper pixel point by the temperature value of the lower pixel point by 2. These two parameters provide information about the temperature change rate of the local area, which is a key indicator for measuring the non-uniformity of the temperature space distribution. These information is very important for identifying temperature mutation area (abnormal hot or cold spot). The temperature change direction helps to understand the heat flow direction or the main trend of heat transfer in the temperature field.

[0108] In step S14, neighborhood analysis is performed according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight.

[0109] In an embodiment, the temperature distribution image is boundary filled to obtain a filled temperature image;

[0110] According to the preset neighborhood rule and the filled temperature image, variance calculation is performed to obtain a neighborhood variance;

[0111] According to the initial temperature difference, gradient calculation is performed to obtain a temperature gradient:

[0112] The weight is calculated by the following formula:

[0113]

[0114] wherein, represents the adaptive weight of a pixel , represents the temperature gradient of a pixel , represents the neighborhood variance.

[0115] It is worth mentioning that the purpose of boundary padding is to avoid the problem of inaccurate calculation caused by the lack of neighborhood information of edge pixels in subsequent processing. By padding the original temperature distribution image, a new padded temperature image is created. The method adopted is to copy the edge pixel value to expand the image boundary, ensuring that each pixel has complete neighborhood information for subsequent calculation.

[0116] It is worth mentioning that the purpose of neighborhood variance calculation is to evaluate the degree of temperature value change in the local area, providing a basis for the calculation of adaptive weight. The preset neighborhood rule is to define a fixed neighborhood size, which is set to 3x3 in this method, to determine the reference area around each pixel point. The variance is calculated based on the filled temperature image, and the variance of the temperature value in the specified neighborhood of each pixel point is calculated. The variance reflects the fluctuation of the temperature value in the region, and the larger the variance, the more intense the temperature change, and vice versa, indicating that the temperature is relatively uniform and stable.

[0117] It is worth mentioning that the purpose of temperature gradient calculation is to quantify the speed and direction of temperature change with spatial position, and to identify areas with sharp temperature changes. Among them, the neighborhood variance acts as a regulator to determine the speed of weight decay. The larger the neighborhood variance value means a more gentle weight decay curve, allowing more high-frequency components to pass through; the smaller the neighborhood variance value, the weight will be quickly reduced, suppressing the influence of noise or mutations. This formula adopts an exponential decay form, when the temperature gradient is large (i.e. there is a significant temperature change), the corresponding weight will decrease rapidly, reflecting that the region contains outliers or noise; on the contrary, for areas with relatively flat temperature changes, the weight is close to 1, emphasizing the importance of these areas. The purpose of this is to effectively suppress noise interference while preserving important features, thereby improving the accuracy of the final calibration results.

[0118] In step S15, according to the adaptive weight, the initial temperature difference and the preset standard deviation threshold, the outlier detection is performed to obtain outlier data.

[0119] In one embodiment, the temperature difference is updated according to the adaptive weight and the initial temperature difference to obtain an updated temperature difference.

[0120] The temperature difference size threshold and the temperature difference direction threshold are obtained.

[0121] When the updated horizontal temperature difference or the updated vertical temperature difference of the updated temperature difference is greater than the temperature difference size threshold, the next step is determined, otherwise the next pixel point is subjected to outlier detection.

[0122] When the updated change direction of the updated temperature difference is greater than the temperature difference direction threshold, the next step is determined, otherwise the next pixel point is subjected to outlier detection.

[0123] The update standard deviation is calculated by the following equation:

[0124]

[0125] wherein, denotes the update standard deviation of pixel , n denotes the number of neighborhood pixels, denotes the neighborhood of pixel , and denotes the update horizontal temperature difference of pixel , and denotes the update vertical temperature difference of pixel , and denotes the average of the update horizontal temperature difference within the neighborhood, denotes the average of the update vertical temperature difference within the neighborhood.

[0126] determining the pixel as an abnormal point when the update standard deviation is greater than the standard deviation threshold value;

[0127] wherein the abnormal point data comprises all the abnormal points.

[0128] In one embodiment, the temperature difference update involves horizontal temperature difference and vertical temperature difference. This process multiplies the original temperature difference by the corresponding weight value. The temperature change direction is not changed.

[0129] It is worth mentioning that the temperature difference size threshold value defines a critical value, only when the update horizontal temperature difference or the update vertical temperature difference of a pixel exceeds this threshold value, it will be further considered whether it is an abnormal point. In addition to considering the size of the temperature difference, it is also checked whether the temperature change direction deviates significantly from the normal range, which is also an auxiliary judgment condition. In this method, the temperature difference size threshold value is set to 2°C; the temperature difference direction threshold value is set to 30°.

[0130] It is worth mentioning that the purpose of the preliminary screening of abnormal points is to quickly filter out the pixels that do not obviously meet the abnormal characteristics according to the two dimensions of temperature difference size and direction. If the update horizontal temperature difference or the update vertical temperature difference of a pixel is greater than the temperature difference size threshold value, it will enter the next step of detailed evaluation; at the same time, it is also necessary to check whether the update temperature change direction of the point exceeds the temperature difference direction threshold value. If both conditions are met, the following steps will be continued; otherwise, skip the current pixel and process the next pixel. Used to reduce calculation.

[0131] ​It is worth noting that the purpose of updating the standard deviation is to quantify the "outlier degree" of a pixel relative to its surrounding neighbors. A higher standard deviation means that the temperature variation pattern of this pixel point is significantly different from its neighbors, which is due to the existence of abnormal situations. The calculated updated standard deviation is compared with a preset standard deviation threshold. If the updated standard deviation of a pixel is greater than the standard deviation threshold, it is considered an abnormal point and is recorded as part of the abnormal point data. The standard deviation threshold is set to 2°C. It is worth noting that a smaller standard deviation threshold (such as 1°C) can more sensitively detect abnormal points in a small range, but at the same time, it also increases the risk of false positives; on the contrary, a larger threshold (such as 3°C) reduces the probability of false positives, but misses some minor abnormal situations.

[0132] In step S16, calibration optimization is performed according to the abnormal point data and the attitude angle data, and a calibrated temperature image is output.

[0133] In an embodiment, the abnormal point data is filled with data to obtain an abnormal correction image.

[0134] The calibrated temperature of the calibrated temperature image is calculated by the following formula:

[0135]

[0136] wherein, represents the calibrated temperature of pixel , represents the corrected temperature value of pixel in the abnormal correction image, represents the pitch angle, represents the roll angle; A calibrated temperature image is generated according to the calibrated temperature, and the calibrated temperature image is output.

[0137] In an embodiment, the temperature value of the abnormal point data is replaced with a neighborhood temperature average.

[0138] It is worth noting that the purpose of filling the abnormal point data is to repair or fill in the data of the detected abnormal points to ensure that the accuracy of the overall result is not affected by these abnormal values in subsequent processing. For all identified abnormal points, the average temperature value of their neighborhood pixels is used to replace the original abnormal temperature value. For example, for a pixel marked as an abnormal point, the average temperature value of all normal pixels in its 3x3 neighborhood is used as the new temperature value of the pixel. The purpose of this is to smooth out the abrupt areas in the image and reduce the impact of noise on the calibration process.

[0139]

[0140] ​It is worth noting that the purpose of this formula is to compensate for temperature measurement errors caused by changes in device posture, i.e., changes in the tilt angle. When the device is not in a horizontal state, the directly measured temperature distribution map will appear distorted. By dividing by the cosine values in both directions, the tilt in a two-dimensional plane is considered, and the error caused by the difference in viewing angle can be corrected to restore the true temperature distribution of the object.

[0141] In summary, the present application discloses an intelligent calibration method for an infrared thermal imager, aiming to solve the measurement error problem existing in the prior art. The method first acquires temperature distribution images and acceleration data to lay the foundation for subsequent processing. Based on these data, further posture angle calculation is performed to obtain the device's posture angle data. This process particularly considers the data deviation caused by the difference in viewing angle when the device measures at different angles, thereby improving the calibration accuracy.

[0142] Specifically, during the posture angle calculation phase, the three-dimensional acceleration values provided by the three-axis accelerometer are used to accurately calculate the pitch angle and roll angle using a specific formula. This not only simplifies the system design, but also enhances the accuracy of the attitude estimation, enabling the infrared thermal imager to automatically adjust its viewing angle and perform image correction. Next, in the temperature difference calculation link, the horizontal and vertical temperature differences between each pixel point and its adjacent pixels are quantified, and the temperature change direction is determined, providing key information about the local area temperature change rate. This is crucial for identifying temperature mutation areas (abnormal hot or cold spots), as this information helps understand the heat flow direction or the main trend of heat transfer in the temperature field.

[0143] Subsequently, the method combines the temperature distribution image with the initial temperature difference to perform neighborhood analysis to obtain adaptive weights. This step first performs boundary filling on the temperature distribution image to avoid edge effects affecting the accuracy of the overall temperature field analysis. Then, according to the preset neighborhood rules, variance calculation is performed to obtain the neighborhood variance. In addition, based on the initial temperature difference, gradient calculation is performed to obtain the temperature gradient of each pixel point, and finally the adaptive weight is calculated by formula. This strategy emphasizes the importance of areas with smooth temperature changes while suppressing the influence of areas with rapid changes or noise interference, improving the overall calibration effect.

[0144] As a key step in the entire calibration process, the abnormal point detection first multiplies the original temperature difference by the corresponding adaptive weight, and then filters out potential abnormal points according to the set temperature difference size and direction threshold. For each candidate abnormal point, the final determination of whether it belongs to an abnormal point is made by calculating the updated standard deviation and comparing it with the preset standard deviation threshold. Once all abnormal points are identified, the average temperature value of the neighboring pixels is used to replace the original abnormal temperature value to repair the mutation area in the image. This process effectively reduces the influence of abnormal values on the quality of the final temperature image, and improves the accuracy and reliability of the temperature distribution.

[0145] Finally, considering the data distortion caused by device tilt, the corrected abnormal point data is combined with the attitude angle data to optimize the calibration and generate a calibrated temperature image that accurately reflects the real temperature distribution of the object. By dividing by the cosine values in both directions, considering the tilt in the two-dimensional plane, the error caused by different viewing angles can be corrected, and the real temperature distribution of the object can be restored. This method not only improves the automation level of the calibration process, but also enhances the adaptability to different environmental conditions, ensuring the accuracy of temperature measurement under different attitudes. In summary, the intelligent calibration method of the infrared thermal imager proposed in the present application significantly improves the ability of the infrared thermal imager to maintain long-term stability and provide accurate temperature measurement results in complex working environments by improving the accuracy of intelligent calibration.

[0146] Referring to Figure 2 The second embodiment of the present application provides an intelligent calibration system for an infrared thermal imager, comprising:

[0147] A data acquisition module for acquiring temperature distribution images and acceleration data;

[0148] An attitude analysis module for calculating attitude angles based on the acceleration data to obtain attitude angle data;

[0149] A temperature difference calculation module for calculating temperature differences based on the temperature distribution images to obtain initial temperature differences;

[0150] A weight calculation module for analyzing the neighborhood based on the temperature distribution images and the initial temperature differences to obtain adaptive weights;

[0151] An abnormality detection module for detecting abnormal points based on the adaptive weights, the initial temperature differences, and a preset standard deviation threshold to obtain abnormal point data;

[0152] A temperature calibration module for optimizing calibration based on the abnormal point data and the attitude angle data to output a calibrated temperature image.

[0153] Preferably, the data acquisition module is configured to:

[0154] Obtaining a temperature distribution image and acceleration data.

[0155] Preferably, the attitude analysis module is configured to calculate an attitude angle based on the acceleration data, and obtain attitude angle data, including:

[0156] The pitch angle is calculated by the following formula:

[0157]

[0158] The roll angle is calculated by the following formula:

[0159]

[0160] wherein, represents the pitch angle, represents the roll angle, represents an acceleration component of the axis, represents an acceleration component of the axis, represents an acceleration component of the axis;

[0161] The attitude angle data includes the pitch angle and the roll angle.

[0162] Preferably, the temperature difference calculation module is configured to:

[0163] Calculate a temperature difference based on the temperature distribution image, and obtain an initial temperature difference, including:

[0164] The horizontal temperature difference is calculated by the following formula:

[0165]

[0166] wherein, represents a horizontal temperature difference of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel

[0167] The vertical temperature difference is calculated by the following formula:

[0168]

[0169] wherein, represents a vertical temperature difference of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel represents a temperature value of a pixel

[0170] The temperature change direction is calculated by the following formula:

[0171]

[0172] wherein, represents the temperature change direction of a pixel .

[0173] wherein the initial temperature difference comprises the temperature change direction, the horizontal temperature difference and the vertical temperature difference.

[0174] Preferably, the weight calculation module is configured to:

[0175] perform neighborhood analysis according to the temperature distribution image and the initial temperature difference to obtain an adaptive weight, comprising:

[0176] perform boundary filling on the temperature distribution image to obtain a filled temperature image;

[0177] perform variance calculation according to a preset neighborhood rule and the filled temperature image to obtain a neighborhood variance;

[0178] perform gradient calculation according to the initial temperature difference to obtain a temperature gradient:

[0179] perform weight calculation by the following formula:

[0180]

[0181] wherein, represents the adaptive weight of a pixel , represents the temperature gradient of a pixel , represents the neighborhood variance.

[0182] Preferably, the anomaly detection module is configured to:

[0183] perform anomaly point detection according to the adaptive weight, the initial temperature difference and a preset standard deviation threshold to obtain anomaly point data, comprising:

[0184] perform temperature difference updating according to the adaptive weight and the initial temperature difference to obtain an updated temperature difference;

[0185] obtain a temperature difference size threshold and a temperature difference direction threshold;

[0186] when an updated horizontal temperature difference or an updated vertical temperature difference of the updated temperature difference is greater than the temperature difference size threshold, perform next step determination, otherwise perform anomaly point detection on a next pixel point;

[0187] When the update change direction of the update temperature difference is greater than the temperature difference direction threshold, the next step is determined, otherwise the next pixel point is detected as an abnormal point;

[0188] The update standard deviation is calculated by the following formula:

[0189]

[0190] wherein, represents the update standard deviation of the pixel , represents the number of neighborhood pixels, represents the neighborhood of the pixel , represents the update horizontal temperature difference of the pixel , represents the update vertical temperature difference of the pixel , represents the average of the update horizontal temperature difference in the neighborhood,

[0191] When the update standard deviation is greater than the standard deviation threshold, the pixel point is determined as an abnormal point.

[0192] The abnormal point data includes all the abnormal points.

[0193] Preferably, the temperature calibration module is configured to:

[0194] According to the abnormal point data and the attitude angle data, a calibration optimization is performed to output a calibration temperature image, including:

[0195] The abnormal point data is filled with data to obtain an abnormal correction image, including:

[0196] The temperature value of the abnormal point data is replaced by the neighborhood temperature average value.

[0197] The calibration temperature of the calibration temperature image is calculated by the following formula:

[0198]

[0199] wherein, represents the calibration temperature of the pixel , represents the correction temperature value of the pixel in the abnormal correction image, represents the pitch angle, represents the roll angle.

[0200] According to the calibration temperature, a calibration temperature image is generated and output. ​​​​​​

[0201] It should be noted that the intelligent calibration system of the infrared thermal imager provided by the embodiments of the present application is used to execute all process steps of the intelligent calibration method of the infrared thermal imager provided by the above embodiments, and the working principles and beneficial effects of the two are one-to-one correspondence, thus not being described again.

[0202] The embodiments of the present application further provide an electronic device. The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, for example, a data acquisition program. The processor implements the steps in the above various embodiments of the intelligent calibration method of the infrared thermal imager when executing the computer program, for example Figure 1 The step S11 shown. Alternatively, the processor implements the functions of the modules / units in the above various device embodiments when executing the computer program, for example, a data acquisition module.

[0203] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0204] The electronic device can be a desktop computer, a notebook, a palm computer, and a smart tablet, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above components are only examples of the electronic device and do not constitute a limitation on the electronic device, and can include more or fewer components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0205] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the electronic device, and connects all parts of the electronic device through various interfaces and lines.

[0206] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required by a function (such as a sound playing function, an image playing function, etc.), and the like; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), and the like. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0207] The modules / units integrated in the electronic device can be stored in a computer readable storage medium if they are realized in the form of software function units and sold or used as independent products. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0208] It should be noted that the apparatus embodiments described above are merely illustrative, and the units described as separate units can or can not be physically separate, and the units displayed as units can or can not be physical units, i.e. can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, the connection relationship between the modules in the apparatus embodiment provided by the present application indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0209] The above specific embodiments further illustrate the purpose, technical scheme and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An intelligent calibration method for an infrared thermal imager, characterized in that: include: Acquire temperature distribution images and acceleration data; Calculating the attitude angle according to the acceleration data to obtain attitude angle data; Calculating the temperature difference based on the temperature distribution image to obtain an initial temperature difference; Performing neighborhood analysis based on the temperature distribution image and the initial temperature difference to obtain an adaptive weight; Perform outlier detection based on the adaptive weight, the initial temperature difference, and a preset standard deviation threshold to obtain outlier data; Performing calibration optimization based on the abnormal point data and the attitude angle data, and outputting a calibrated temperature image; The performing of outlier detection according to the adaptive weight, the initial temperature difference, and a preset standard deviation threshold to obtain outlier data includes: Performing a temperature difference update according to the adaptive weight and the initial temperature difference to obtain an updated temperature difference; Get the temperature difference magnitude threshold and temperature difference direction threshold; When the updated horizontal temperature difference or the updated vertical temperature difference of the updated temperature difference is greater than the temperature difference threshold, the next step is determined; otherwise, an abnormal point detection is performed on the next pixel; When the updated change direction of the updated temperature difference is greater than the temperature difference direction threshold, the next step is determined; otherwise, an abnormal point detection is performed on the next pixel; The updated standard deviation is calculated using the following formula: in, Represents pixels The updated standard deviation of represents the number of neighborhood pixels, Represents pixels Neighborhood, Represents pixels The updated horizontal temperature difference, Represents pixels The updated vertical temperature difference, represents the average of the updated horizontal temperature differences within the neighborhood, represents the average of the updated vertical temperature differences within the neighborhood; When the updated standard deviation is greater than the standard deviation threshold, the pixel is determined to be an abnormal point; Wherein, the outlier data includes all the outliers; The step of performing calibration optimization based on the abnormal point data and the attitude angle data and outputting a calibration temperature image includes: Performing data filling on the abnormal point data to obtain an abnormality corrected image; The calibration temperature of the calibration temperature image is calculated by the following formula: in, Represents pixels The calibration temperature, Indicates the pixels in the abnormally corrected image The corrected temperature value, represents the pitch angle, represents the roll angle; A calibration temperature image is generated according to the calibration temperature, and the calibration temperature image is output.

2. The intelligent calibration method for infrared thermal imagers according to claim 1, characterized in that: The step of calculating the attitude angle according to the acceleration data to obtain attitude angle data includes: The pitch angle is calculated using the following formula: The roll angle is calculated using the following formula: in, represents the pitch angle, represents the roll angle, express The acceleration components of the axes, express The acceleration components of the axes, express The acceleration components of the axes; The attitude angle data includes the pitch angle and the roll angle.

3. The intelligent calibration method for infrared thermal imager according to claim 1, characterized in that: The step of calculating the temperature difference according to the temperature distribution image to obtain the initial temperature difference includes: The horizontal temperature difference is calculated by the following formula: in, Represents pixels The horizontal temperature difference, Represents pixels The temperature value, Represents pixels Temperature value; The vertical temperature difference is calculated using the following formula: in, Represents pixels The vertical temperature difference, Represents pixels The temperature value, Represents pixels Temperature value; The direction of temperature change is calculated using the following formula: in, Represents pixels The direction of temperature change; The initial temperature difference includes the temperature change direction, the horizontal temperature difference and the vertical temperature difference.

4. The intelligent calibration method for infrared thermal imagers according to claim 1, characterized in that: The performing neighborhood analysis based on the temperature distribution image and the initial temperature difference to obtain an adaptive weight includes: Filling the boundaries of the temperature distribution image to obtain a filled temperature image; Performing variance calculation based on a preset neighborhood rule and the filling temperature image to obtain a neighborhood variance; The gradient is calculated based on the initial temperature difference to obtain the temperature gradient: The weight is calculated using the following formula: in, Represents pixels The adaptive weight of Represents pixels The temperature gradient, represents the neighborhood variance.

5. The intelligent calibration method for infrared thermal imager according to claim 1, characterized in that: The step of performing calibration optimization based on the abnormal point data and the attitude angle data and outputting a calibration temperature image includes: Performing data filling on the abnormal point data to obtain an abnormality corrected image; The calibration temperature of the calibration temperature image is calculated by the following formula: in, Represents pixels The calibration temperature, Indicates the pixels in the abnormally corrected image The corrected temperature value, represents the pitch angle, represents the roll angle; A calibration temperature image is generated according to the calibration temperature, and the calibration temperature image is output.

6. The intelligent calibration method for infrared thermal imager according to claim 5, characterized in that: The step of performing data filling on the outlier point data to obtain an outlier corrected image includes: The temperature value of the abnormal point data is replaced by the average temperature of the neighborhood.

7. An intelligent calibration system for an infrared thermal imager, used to implement the intelligent calibration method for an infrared thermal imager according to any one of claims 1 to 6, characterized in that: include: A data acquisition module, used to acquire temperature distribution images and acceleration data; A posture analysis module, configured to calculate the posture angle according to the acceleration data to obtain posture angle data; a temperature difference calculation module, configured to calculate the temperature difference according to the temperature distribution image to obtain an initial temperature difference; a weight calculation module, configured to perform neighborhood analysis based on the temperature distribution image and the initial temperature difference to obtain an adaptive weight; an anomaly detection module, configured to perform anomaly detection based on the adaptive weight, the initial temperature difference, and a preset standard deviation threshold to obtain anomaly data; The temperature calibration module is used to perform calibration optimization based on the abnormal point data and the attitude angle data, and output a calibrated temperature image.

8. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method implements the intelligent calibration method for the infrared thermal imager according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the intelligent calibration method for the infrared thermal imager according to any one of claims 1 to 6.

Images