Dynamic range self-adaptive adjusting system and adjusting method of thermal infrared imager

Abstract

The invention discloses a dynamic range self-adaptive adjusting system and adjusting method of a thermal infrared imager, and relates to the technical field of thermal infrared imager control, and the dynamic range self-adaptive adjusting system comprises a self-adaptive dynamic range adjusting module, an intelligent thermal noise filtering module and a self-adaptive control module. The self-adaptive dynamic range adjusting module comprises a temperature monitoring unit, an image partitioning unit and a contrast enhancement unit. According to the dynamic range self-adaptive adjusting system and method of the thermal infrared imager, the imaging definition can be effectively improved through self-adaptive adjustment of the dynamic range and intelligent filtering of thermal noise, image details are more prominent, and the system and method are suitable for accurate monitoring in a low-contrast or high-temperature-difference environment; the adaptive control algorithm can analyze and optimize various parameters in real time, so that the imaging requirements of various complex environments are met, and the dependence on the operation environment is reduced; manual adjustment is not needed, the operation complexity is reduced, the stability of an imaging result is improved, and the method is suitable for long-term monitoring and unattended application.

Description

Technical Field

[0001] The invention relates to the field of infrared thermal imager control technology, and in particular to a dynamic range adaptive adjustment system and an adjustment method for an infrared thermal imager. Background Art

[0002] Existing infrared thermal imagers are widely used in a variety of industrial and inspection scenarios, such as equipment fault detection, building energy management, and night vision imaging. However, due to the significant differences in ambient and object temperatures, different scenarios have different requirements for the dynamic range of infrared imaging, and ordinary thermal imagers are usually unable to adapt to environmental requirements.

[0003] In addition, the impact of thermal noise in complex scenes will reduce the readability of the image, leading to an increased risk of misjudgment. Therefore, the existing technology still needs to be improved in terms of efficient image processing, automatic adaptive dynamic adjustment, and thermal noise filtering. Summary of the invention

[0004] In order to solve the above-mentioned technical problems of infrared thermal imagers, the present invention provides a dynamic range adaptive adjustment system and adjustment method of infrared thermal imagers. The following technical solutions are adopted: The dynamic range adaptive adjustment system of the infrared thermal imager includes an adaptive dynamic range adjustment module, an intelligent thermal noise filtering module and an adaptive control module; The adaptive dynamic range adjustment module includes a temperature monitoring unit, an image partitioning unit and a contrast enhancement unit. The temperature monitoring unit collects temperature distribution information of the current scene and generates thermal map data. The image partitioning unit is connected in communication with the temperature monitoring unit. The thermal map is divided into a plurality of independent regions based on the temperature distribution information data through a region segmentation algorithm. The contrast enhancement unit is connected in communication with the image partitioning unit, receives data of each independent region and performs contrast adjustment. The intelligent thermal noise filtering module includes a noise detection unit, a background modeling unit and a dynamic filtering unit. The noise detection unit is connected to the contrast enhancement unit for communication, identifies and marks low signal-to-noise ratio areas in the image, and the background modeling unit is connected to the noise detection unit for communication. The noise data marked by the noise detection unit is transmitted to the background modeling unit, and the background modeling unit generates a noise feature model for a specific scene. The dynamic filtering unit is connected to the background modeling unit for communication, and the dynamic filtering unit sets filtering parameters based on the deviation value of the noise feature model, and performs filtering operations on each independent area respectively. The adaptive control module includes a data analysis module and a parameter adaptive adjustment module. The data analysis module is in communication with the temperature monitoring unit and the intelligent thermal noise filtering module, and uses a machine learning model to perform in-depth analysis on the input temperature and noise data to determine the temperature range and noise characteristics of different areas; The parameter adaptive adjustment module is respectively connected to the data analysis module, the contrast enhancement unit and the dynamic filtering unit for communication. The parameter adaptive adjustment module dynamically adjusts various parameters of the contrast enhancement unit and the dynamic filtering unit in combination with the temperature and noise characteristics output by the data analysis module.

[0005] By adopting the above technical solutions, the adaptive dynamic range adjustment module aims to dynamically adjust the imaging contrast and detail presentation of the infrared thermal imager according to the requirements of different temperature environments; The specific temperature monitoring unit samples the temperature of the image at a preset time interval (such as every second) and transmits the temperature data to the image partitioning unit. The temperature monitoring unit can also predict potential high temperature areas based on historical data and temperature change trends to provide a reference for dynamic range adjustment.

[0006] The image partitioning unit can adjust the dynamic range of each area separately through the area segmentation algorithm to ensure the best contrast in each area and achieve a balance between details and global clarity.

[0007] The contrast enhancement unit adjusts the contrast of each partition so that the area with more significant temperature difference shows more details. The specific contrast enhancement unit uses adaptive histogram equalization or Laplace enhancement algorithm to enhance the contrast, especially enhancing the details in the temperature extreme value area.

[0008] After the temperature monitoring unit detects the temperature data, the image partitioning unit divides the area. The contrast enhancement unit receives the data of each partition and performs contrast adjustment, and then synthesizes the final thermal image; The intelligent thermal noise filtering module intelligently identifies and removes invalid noise in low-contrast or thermal noise environments to improve image quality. The noise detection unit identifies possible noise areas by analyzing the noise characteristics and temperature fluctuations of the image. By combining spatial filtering with temporal filtering, the short-term and long-term changes in pixel temperature are detected to distinguish normal temperature changes from random noise. The noise detection unit uses Gaussian filtering, variance analysis and other techniques to identify and mark low signal-to-noise ratio areas.

[0009] The background modeling unit generates a background model of inefficient thermal noise to facilitate effective filtering in subsequent images. The background modeling unit performs superposition analysis on multiple frames of images to gradually extract constant or low-frequency thermal noise patterns; The dynamic filtering unit dynamically filters invalid thermal noise data in the image according to the noise feature model of the background modeling unit. Dynamic filtering is performed using the noise features generated by background modeling. The dynamic filtering unit sets filtering parameters based on the deviation value of the noise feature model, and does not filter areas that obviously deviate from the noise features, thereby ensuring that the true temperature signal will not be deleted by mistake.

[0010] The adaptive control module dynamically adjusts and optimizes control parameters by analyzing the temperature distribution and noise characteristics of the imaging scene to ensure optimal image clarity.

[0011] Optionally, the temperature monitoring unit includes a temperature sensor, an analog-to-digital converter, a temperature data storage device and a microprocessor, wherein the temperature sensor collects temperature distribution information of the current scene, the analog-to-digital converter is communicated with the temperature sensor, the temperature data storage device is connected to the digital output end of the analog-to-digital converter, and the microprocessor is communicated with the temperature data storage device to analyze the temperature distribution information, output thermal map data and store it in the temperature data storage device.

[0012] By adopting the above technical solution, the temperature sensor can be an infrared sensor, which measures the temperature by detecting infrared radiation emitted by the object, and the microprocessor can form an initial thermal map based on the data of the temperature sensor.

[0013] Optionally, the image partitioning unit includes an image buffer and an image analysis chip, wherein the image buffer is connected to the temperature data storage device for communicating and interacting with the thermal map data, and the image analysis chip divides the thermal map into several independent regions based on the temperature distribution information data through a region segmentation algorithm.

[0014] By adopting the above technical solutions, the image analysis chip can perform regional segmentation algorithms, such as the region growing method based on the temperature gradient to divide the parts with significant temperature differences into independent regions, paying special attention to the areas with higher temperatures or more dramatic changes. The dynamic range of each region can be adjusted separately to ensure the best contrast in each region and achieve a balance between details and global clarity.

[0015] Optionally, the contrast enhancement unit, the noise detection unit and the dynamic filtering unit are all implemented based on an image analysis processor.

[0016] Optionally, the background modeling unit includes a data storage device and a data analysis chip, wherein the data storage device is communicatively connected to the noise detection unit, and the data analysis chip constructs a noise characteristic model based on the marked noise data.

[0017] By adopting the above technical solutions, the data analysis chip can generate a noise feature model for a specific scene after analyzing multiple heat map frames based on motion estimation and background difference. The background modeling unit can be automatically updated when the environment changes to maintain the accuracy of the noise feature model. Optionally, the parameter adaptive adjustment module includes a microcontroller, a field programmable gate array and a parameter adjustment instruction buffer, the microcontroller is communicatively connected to the data analysis module, the field programmable gate array is communicatively connected to the microcontroller, and is used to implement an adaptive algorithm and generate real-time parameter adjustment instructions, the parameter adjustment instruction buffer communicates with the field programmable gate array to exchange real-time parameter adjustment instructions, and is respectively communicatively connected to the contrast enhancement unit and the dynamic filtering unit to exchange real-time parameter adjustment instructions.

[0018] By adopting the above technical solutions, the microcontroller is the core of the parameter adaptive adjustment module, responsible for executing algorithms, processing data, and adjusting parameters according to preset conditions or real-time feedback; the field programmable gate array (FPGA) provides a high degree of customizability and parallel processing capabilities, which is suitable for implementing complex adaptive algorithms and real-time parameter adjustment.

[0019] The method for adaptively adjusting the dynamic range of an infrared thermal imager adopts an adaptively adjusting system of the dynamic range of an infrared thermal imager to adaptively adjust the infrared thermal imager, and includes the following steps: Step 1: The temperature monitoring unit samples the temperature at a preset time interval, generates thermal map data, and transmits the thermal map data to the image partitioning unit; Step 2, the image partitioning unit divides the thermal image into several independent regions based on the temperature distribution information data through a region segmentation algorithm; Step 3: The contrast enhancement unit receives the data of each independent region, uses an adaptive histogram equalization or Laplace enhancement algorithm to enhance the contrast of each independent region, and merges the independent regions after contrast enhancement into a heat map; Step 4: The noise detection unit uses Gaussian filtering or variance analysis technology to identify and mark low signal-to-noise ratio areas; Step 5, the noise data marked by the noise detection unit is transmitted to the background modeling unit, and the background modeling unit establishes a noise feature model including noise features after multiple frame analyses based on motion estimation and background difference; Step 6: The dynamic filtering unit sets filtering parameters based on the deviation value of the noise feature model, and performs filtering operations on each independent area to achieve intelligent noise filtering; Step 7, the microcontroller analyzes the input temperature and noise data to determine the temperature range and noise characteristics of different independent areas, and the field programmable gate array generates real-time instructions for optimizing the parameter values ​​of the contrast enhancement unit and the dynamic filtering unit, and interacts with the contrast enhancement unit and the dynamic filtering unit through the parameter adjustment instruction buffer to issue real-time parameter adjustment instructions. The contrast enhancement unit and the dynamic filtering unit execute the real-time parameter adjustment instructions to achieve adaptive adjustment of the dynamic range.

[0020] Optionally, in step 6, the dynamic filtering unit does not filter the independent region that reaches the deviation noise feature threshold.

[0021] By adopting the above technical solution, dynamic filtering is performed using the noise characteristics generated by background modeling. The dynamic filtering unit sets the filtering parameters based on the deviation value of the noise characteristic model, and does not filter the areas that obviously deviate from the noise characteristics, thereby ensuring that the real temperature signal will not be mistakenly deleted.

[0022] The determination of the region that obviously deviates from the noise feature can be realized by setting a threshold for deviation from the noise feature. Specifically, one or more thresholds are set according to the statistical characteristics of the noise. These thresholds can be fixed or adaptive, depending on the dynamic characteristics of the noise.

[0023] For example, a standard deviation threshold can be set, and if the value of a data point is outside the range of the mean plus or minus a certain number of standard deviations, it is considered an outlier.

[0024] Optionally, in step 7, the microcontroller uses a machine learning model to perform in-depth analysis on the input temperature and noise data.

[0025] Optionally, in step 7, the field programmable gate array generates in real time instructions for optimizing parameter values ​​of the contrast enhancement unit and the dynamic filtering unit based on a gradient descent method and a feedback adjustment mechanism.

[0026] In summary, the present invention includes at least one of the following beneficial technical effects: The present invention can provide a dynamic range adaptive adjustment system and adjustment method for an infrared thermal imager, which can effectively improve the imaging clarity and make the image details more prominent by adaptively adjusting the dynamic range and intelligently filtering thermal noise, and is suitable for accurate monitoring in low-contrast or high-temperature difference environments; Adaptive control algorithms can analyze and optimize various parameters in real time to meet the imaging needs of various complex environments, so that infrared thermal imagers can be widely used in different scenarios and reduce dependence on the operating environment; The system will automatically optimize the imaging effect without manual adjustment, which not only reduces the complexity of operation but also improves the stability of imaging results, making it suitable for long-term monitoring and unattended applications. Through dynamic range adjustment and intelligent thermal noise filtering functions, it solves the problems of image clarity and automatic adaptability in complex environments, and is suitable for a wide range of industrial and environmental monitoring applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the connection principle of the electrical components of the dynamic range adaptive adjustment system of the infrared thermal imager of the present invention; Figure 2It is a flow chart of the method for adaptively adjusting the dynamic range of the infrared thermal imager of the present invention.

[0028] Explanation of the accompanying drawings: 1. temperature monitoring unit; 11. temperature sensor; 12. analog-to-digital converter; 13. temperature data storage device; 14. microprocessor; 2. image partitioning unit; 21. image buffer; 22. image analysis chip; 3. contrast enhancement unit; 41. noise detection unit; 42. background modeling unit; 421. data storage device; 422. data analysis chip; 43. dynamic filtering unit; 51. data analysis module; 52. parameter adaptive adjustment module; 521. microcontroller; 522. field programmable gate array; 523. parameter adjustment instruction buffer. DETAILED DESCRIPTION

[0029] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0030] The embodiment of the invention discloses a dynamic range adaptive adjustment system and an adjustment method of an infrared thermal imager.

[0031] Reference Figure 1 and Figure 2 , Embodiment 1, a dynamic range adaptive adjustment system for an infrared thermal imager, comprising an adaptive dynamic range adjustment module, an intelligent thermal noise filtering module and an adaptive control module; The adaptive dynamic range adjustment module includes a temperature monitoring unit 1, an image partitioning unit 2 and a contrast enhancement unit 3. The temperature monitoring unit 1 collects the temperature distribution information of the current scene and generates thermal map data. The image partitioning unit 2 is connected in communication with the temperature monitoring unit 1. Through the region segmentation algorithm, the thermal map is divided into several independent regions based on the temperature distribution information data. The contrast enhancement unit 3 is connected in communication with the image partitioning unit 2, receives the data of each independent region and performs contrast adjustment. The intelligent thermal noise filtering module includes a noise detection unit 41, a background modeling unit 42 and a dynamic filtering unit 43. The noise detection unit 41 is connected to the contrast enhancement unit 3 for communication, identifies and marks low signal-to-noise ratio areas in the image, and the background modeling unit 42 is connected to the noise detection unit 41 for communication. The noise data marked by the noise detection unit 41 is transmitted to the background modeling unit 42, and the background modeling unit 42 generates a noise feature model for a specific scene. The dynamic filtering unit 43 is connected to the background modeling unit 42 for communication, and the dynamic filtering unit 43 sets filtering parameters based on the deviation value of the noise feature model, and performs filtering operations on each independent area respectively. The adaptive control module includes a data analysis module 51 and a parameter adaptive adjustment module 52. The data analysis module 51 is in communication with the temperature monitoring unit 1 and the intelligent thermal noise filtering module, and uses a machine learning model to perform in-depth analysis on the input temperature and noise data to determine the temperature range and noise characteristics of different areas; The parameter adaptive adjustment module 52 is respectively connected to the data analysis module 51, the contrast enhancement unit 3 and the dynamic filtering unit 43 for communication. The parameter adaptive adjustment module 52 dynamically adjusts the parameters of the contrast enhancement unit 3 and the dynamic filtering unit 43 in combination with the temperature and noise characteristics output by the data analysis module 51.

[0032] The adaptive dynamic range adjustment module is designed to dynamically adjust the imaging contrast and detail presentation of the infrared thermal imager according to the needs of different temperature environments; The specific temperature monitoring unit 1 samples the temperature of the image at a preset time interval (such as every second) and transmits the temperature data to the image partitioning unit. The temperature monitoring unit can also predict potential high temperature areas based on historical data and temperature change trends to provide a reference for dynamic range adjustment.

[0033] The image partitioning unit 2 can adjust the dynamic range of each region separately through the region segmentation algorithm to ensure the best contrast of each region and achieve a balance between details and global clarity.

[0034] The contrast enhancement unit 3 adjusts the contrast of each partition so that the area with more significant temperature difference shows more details. Specifically, the contrast enhancement unit 3 uses adaptive histogram equalization or Laplace enhancement algorithm to enhance the contrast, especially to enhance the details in the temperature extreme value area.

[0035] After the temperature monitoring unit 1 detects the temperature data, the image partitioning unit 2 performs region division. The contrast enhancement unit receives the data of each partition and performs contrast adjustment, and then synthesizes the final thermal image; The intelligent thermal noise filtering module intelligently identifies and removes invalid noise in low-contrast or thermal noise environments to improve image quality. The noise detection unit 41 identifies possible noise areas by analyzing the noise characteristics and temperature fluctuations of the image. By combining spatial filtering with temporal filtering, the short-term and long-term changes in pixel temperature are detected to distinguish normal temperature changes from random noise. The noise detection unit uses Gaussian filtering, variance analysis and other techniques to identify and mark low signal-to-noise ratio areas.

[0036] The background modeling unit 42 generates a background model of inefficient thermal noise to facilitate effective filtering in subsequent images. The background modeling unit 42 performs superposition analysis on multiple frames of images to gradually extract constant or low-frequency thermal noise patterns; The dynamic filtering unit 43 dynamically filters the invalid thermal noise data in the image according to the noise feature model of the background modeling unit. Dynamic filtering is performed using the noise features generated by background modeling. The dynamic filtering unit sets the filtering parameters based on the deviation value of the noise feature model, and does not filter the areas that obviously deviate from the noise features, thereby ensuring that the true temperature signal will not be deleted by mistake.

[0037] The adaptive control module dynamically adjusts and optimizes control parameters by analyzing the temperature distribution and noise characteristics of the imaging scene to ensure optimal image clarity.

[0038] Embodiment 2, the temperature monitoring unit 1 includes a temperature sensor 11, an analog-to-digital converter 12, a temperature data storage 13 and a microprocessor 14, the temperature sensor 11 collects the temperature distribution information of the current scene, the analog-to-digital converter 12 is communicated with the temperature sensor 11, the temperature data storage 13 is connected to the digital output end of the analog-to-digital converter 12, the microprocessor 14 is communicated with the temperature data storage 13, analyzes the temperature distribution information, outputs the thermal map data and stores it in the temperature data storage 13.

[0039] The temperature sensor 11 may be an infrared sensor that measures temperature by detecting infrared radiation emitted by an object. The microprocessor 14 may form an initial thermal map based on data from the temperature sensor 11 .

[0040] In embodiment 3, the image partitioning unit 2 includes an image buffer 21 and an image analysis chip 22. The image buffer 21 communicates with the temperature data storage 13 to exchange thermal map data. The image analysis chip 22 divides the thermal map into several independent regions based on the temperature distribution information data through a region segmentation algorithm.

[0041] The image analysis chip 22 can execute a region segmentation algorithm, such as a region growing method based on a temperature gradient to divide the portion with a significant temperature difference into independent regions, with special attention paid to regions with higher temperatures or more dramatic changes. The dynamic range of each region can be adjusted separately to ensure the best contrast of each region and achieve a balance between details and global clarity.

[0042] In the fourth embodiment, the contrast enhancement unit 3, the noise detection unit 41 and the dynamic filtering unit 43 are all implemented based on an image analysis processor.

[0043] In Embodiment 5, the background modeling unit 42 includes a data storage 421 and a data analysis chip 422. The data storage 421 is in communication connection with the noise detection unit 41. The data analysis chip 422 constructs a noise characteristic model based on the marked noise data.

[0044] The data analysis chip 422 can generate a noise feature model for a specific scene after analyzing multiple heat map frames based on motion estimation and background difference. The background modeling unit can be automatically updated when the environment changes to maintain the accuracy of the noise feature model. Embodiment 6, the parameter adaptive adjustment module 52 includes a microcontroller 521, a field programmable gate array 522 and a parameter adjustment instruction buffer 523, the microcontroller 521 is communicated with the data analysis module 51, the field programmable gate array 522 is communicated with the microcontroller 521, and is used to implement adaptive algorithms and real-time parameter adjustment instruction generation, the parameter adjustment instruction buffer 523 communicates with the field programmable gate array 522 to exchange real-time parameter adjustment instructions, and is respectively communicated with the contrast enhancement unit 3 and the dynamic filtering unit 43 to exchange real-time parameter adjustment instructions.

[0045] The microcontroller 521 is the core of the parameter adaptive adjustment module, responsible for executing algorithms, processing data, and adjusting parameters according to preset conditions or real-time feedback; the field programmable gate array (FPGA) provides a high degree of customizability and parallel processing capabilities, suitable for implementing complex adaptive algorithms and real-time parameter adjustment.

[0046] Embodiment 7, a method for adaptively adjusting the dynamic range of an infrared thermal imager, using an adaptive dynamic range adjustment system of an infrared thermal imager to adaptively adjust the infrared thermal imager, comprises the following steps: Step 1, the temperature monitoring unit 1 performs temperature sampling at preset time intervals, generates thermal map data, and transmits the thermal map data to the image partitioning unit 2; Step 2, the image partitioning unit 2 divides the thermal image into a number of independent regions based on the temperature distribution information data through a region segmentation algorithm; Step 3, the contrast enhancement unit 3 receives the data of each independent area, uses an adaptive histogram equalization or Laplace enhancement algorithm to enhance the contrast of each independent area, and merges the independent areas after the contrast enhancement into a heat map; Step 4, the noise detection unit 41 uses Gaussian filtering or variance analysis technology to identify and mark the low signal-to-noise ratio area; Step 5, the noise data marked by the noise detection unit 41 is transmitted to the background modeling unit 42, and the background modeling unit 42 establishes a noise feature model including noise features after multiple frame analyses based on motion estimation and background difference; Step 6, the dynamic filtering unit 43 sets filtering parameters based on the deviation value of the noise characteristic model, and performs filtering operations on each independent area to achieve intelligent noise filtering; Step 7, the microcontroller 521 analyzes the input temperature and noise data to determine the temperature range and noise characteristics of different independent areas, and the field programmable gate array 522 generates real-time instructions for optimizing the parameter values ​​of the contrast enhancement unit 3 and the dynamic filtering unit 43, and interacts with the contrast enhancement unit 3 and the dynamic filtering unit 43 through the parameter adjustment instruction buffer 523 to generate real-time parameter adjustment instructions, and the contrast enhancement unit 3 and the dynamic filtering unit 43 execute the real-time parameter adjustment instructions to achieve dynamic range adaptive adjustment.

[0047] In Embodiment 8, in step 6, the dynamic filtering unit 43 does not filter the independent regions that reach the threshold value of deviation from the noise characteristic.

[0048] Dynamic filtering is performed using the noise characteristics generated by background modeling. The dynamic filtering unit sets the filtering parameters based on the deviation value of the noise characteristic model, and does not filter the areas that obviously deviate from the noise characteristics, thereby ensuring that the real temperature signal will not be mistakenly deleted.

[0049] The determination of the region that obviously deviates from the noise feature can be realized by setting a threshold for deviation from the noise feature. Specifically, one or more thresholds are set according to the statistical characteristics of the noise. These thresholds can be fixed or adaptive, depending on the dynamic characteristics of the noise.

[0050] For example, a standard deviation threshold can be set, and if the value of a data point is outside the range of the mean plus or minus a certain number of standard deviations, it is considered an outlier.

[0051] In Example 9, in step 7, the microcontroller 521 uses a machine learning model to perform in-depth analysis on the input temperature and noise data.

[0052] In Example 10, in step 7, the field programmable gate array 522 generates in real time instructions for optimizing the parameter values ​​of the contrast enhancement unit 3 and the dynamic filtering unit 43 based on the gradient descent method and the feedback adjustment mechanism.

[0053] The specific method of the gradient descent method and feedback adjustment mechanism is to first define a cost function, which evaluates the image quality under the current parameter settings, which can be indicators such as contrast level and noise level.

[0054] Calculate the gradient of the cost function with respect to the parameters: The gradient indicates the direction in which the cost function grows fastest. We need to adjust the parameters in the opposite direction of the gradient to reduce the cost function.

[0055] Apply gradient descent to update parameters: Update parameters according to gradient and learning rate.

[0056] Feedback adjustment: Based on the new parameter settings, the cost function is evaluated again. If the cost function does not improve significantly, it may be necessary to adjust the learning rate or stop updating.

[0057] The specific code example is as follows: import numpy as np # Hypothesized cost function for evaluating heatmap quality def cost_function(params, image): # Here we use a simple example. In actual applications, we need to design the cost function according to the specific situation. clarity_score = np.std(image) # Use standard deviation as a measure of contrast level return -clarity_score #The negative sign is because the gradient descent method aims to minimize the cost # Calculate the gradient of the cost function with respect to the parameters def compute_gradient(params, image): # Compute the gradient using numerical differentiation: epsilon=1e-6 grad = np.zeros_like(params) for i in range(len(params)): params_plus = params.copy() params_minus = params.copy() params_plus[i] +=epsilon params_minus[i] -= epsilon grad[i]=(cost_function(params_plus,image)- cost_function(params_minus, image)) / (2 * epsilon) return grad # Gradient descent method to update parameters: Defgradient_descent(params,image,learning_rate, num_iterations): for i in range(num_iterations): grad=compute_gradient(params, image) params -= learning_rate * grad # Initialization parameters initial_params=np.array([0.5, 0.5]) # Assume there are two parameters to be optimized learning_rate=0.01 #Learning rate num_iterations=100 #Number of iterations # Assume that image is the heat map data currently being processed #optimized_params=gradient_descent(initial_params,image, learning_rate, num_iterations).

[0058] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The dynamic range adaptive adjustment system of the infrared thermal imager is characterized by: It includes an adaptive dynamic range adjustment module, an intelligent thermal noise filtering module and an adaptive control module; The adaptive dynamic range adjustment module comprises a temperature monitoring unit (1), an image partitioning unit (2) and a contrast enhancement unit (3), wherein the temperature monitoring unit (1) collects temperature distribution information of a current scene and generates thermal map data, the image partitioning unit (2) is connected in communication with the temperature monitoring unit (1), and divides the thermal map into a plurality of independent regions based on the temperature distribution information data through a region segmentation algorithm, and the contrast enhancement unit (3) is connected in communication with the image partitioning unit (2), receives data of each independent region and performs contrast adjustment; The intelligent thermal noise filtering module comprises a noise detection unit (41), a background modeling unit (42) and a dynamic filtering unit (43); the noise detection unit (41) is connected in communication with the contrast enhancement unit (3) to identify and mark a low signal-to-noise ratio area in an image; the background modeling unit (42) is connected in communication with the noise detection unit (41); the noise data marked by the noise detection unit (41) is transmitted to the background modeling unit (42); and the background modeling unit (42) generates a noise feature model of a specific scene; The dynamic filtering unit (43) is in communication connection with the background modeling unit (42), and the dynamic filtering unit (43) sets filtering parameters based on the deviation value of the noise characteristic model, and performs filtering operations on each independent area respectively; The adaptive control module comprises a data analysis module (51) and a parameter adaptive adjustment module (52); the data analysis module (51) is in communication with the temperature monitoring unit (1) and the intelligent thermal noise filtering module, and uses a machine learning model to perform in-depth analysis on input temperature and noise data to determine temperature ranges and noise characteristics of different regions; The parameter adaptive adjustment module (52) is respectively connected to the data analysis module (51), the contrast enhancement unit (3) and the dynamic filtering unit (43) for communication. The parameter adaptive adjustment module (52) dynamically adjusts various parameters of the contrast enhancement unit (3) and the dynamic filtering unit (43) in combination with the temperature and noise characteristics output by the data analysis module (51).

2. The dynamic range adaptive adjustment system of the infrared thermal imager according to claim 1, characterized in that: The temperature monitoring unit (1) comprises a temperature sensor (11), an analog-to-digital converter (12), a temperature data storage (13) and a microprocessor (14); the temperature sensor (11) collects temperature distribution information of a current scene; the analog-to-digital converter (12) is communicatively connected to the temperature sensor (11); the temperature data storage (13) is communicatively connected to a digital output terminal of the analog-to-digital converter (12); the microprocessor (14) is communicatively connected to the temperature data storage (13), analyzes the temperature distribution information, outputs thermal map data and stores the data in the temperature data storage (13).

3. The dynamic range adaptive adjustment system of the infrared thermal imager according to claim 2, characterized in that: The image partitioning unit (2) comprises an image buffer (21) and an image analysis chip (22); the image buffer (21) is connected in communication with the temperature data storage device (13) to exchange thermal image data; and the image analysis chip (22) divides the thermal image into a plurality of independent regions based on the temperature distribution information data through a region segmentation algorithm.

4. The dynamic range adaptive adjustment system of the infrared thermal imager according to claim 3, characterized in that: The contrast enhancement unit (3), the noise detection unit (41) and the dynamic filtering unit (43) are all implemented based on an image analysis processor.

5. The dynamic range adaptive adjustment system of the infrared thermal imager according to claim 4, characterized in that: The background modeling unit (42) comprises a data storage (421) and a data analysis chip (422), wherein the data storage (421) is communicatively connected to the noise detection unit (41), and the data analysis chip (422) constructs a noise characteristic model based on the marked noise data.

6. The dynamic range adaptive adjustment system of the infrared thermal imager according to claim 5, characterized in that: The parameter adaptive adjustment module (52) comprises a microcontroller (521), a field programmable gate array (522) and a parameter adjustment instruction buffer (523); the microcontroller (521) is communicatively connected to the data analysis module (51); the field programmable gate array (522) is communicatively connected to the microcontroller (521) for realizing an adaptive algorithm and generating real-time parameter adjustment instructions; the parameter adjustment instruction buffer (523) communicates and exchanges real-time parameter adjustment instructions with the field programmable gate array (522), and is respectively communicatively connected to the contrast enhancement unit (3) and the dynamic filtering unit (43) for exchanging real-time parameter adjustment instructions.

7. A method for adaptively adjusting the dynamic range of an infrared thermal imager, characterized in that: Adopting the dynamic range adaptive adjustment system of the infrared thermal imager according to claim 6 to adaptively adjust the infrared thermal imager comprises the following steps: Step 1, the temperature monitoring unit (1) performs temperature sampling at preset time intervals, generates thermal map data, and transmits the thermal map data to the image partitioning unit (2); Step 2, the image partitioning unit (2) divides the thermal image into a plurality of independent regions based on the temperature distribution information data by using a region segmentation algorithm; Step 3, the contrast enhancement unit (3) receives the data of each independent region, uses an adaptive histogram equalization or Laplace enhancement algorithm to enhance the contrast of each independent region, and merges the independent regions after the contrast enhancement into a heat map; Step 4, the noise detection unit (41) uses Gaussian filtering or variance analysis technology to identify and mark the low signal-to-noise ratio area; Step 5, the noise data marked by the noise detection unit (41) is transmitted to the background modeling unit (42), and the background modeling unit (42) establishes a noise feature model including noise features after multiple frame analyses based on motion estimation and background difference; Step 6, the dynamic filtering unit (43) sets filtering parameters based on the deviation value of the noise characteristic model, and performs filtering operations on each independent area to achieve intelligent noise filtering; Step 7, the microcontroller (521) analyzes the input temperature and noise data to determine the temperature range and noise characteristics of different independent areas, the field programmable gate array (522) generates in real time instructions for optimizing the parameter values ​​of the contrast enhancement unit (3) and the dynamic filtering unit (43), and interacts with the contrast enhancement unit (3) and the dynamic filtering unit (43) through the parameter adjustment instruction buffer (523) to exchange real-time parameter adjustment instructions, and the contrast enhancement unit (3) and the dynamic filtering unit (43) execute the real-time parameter adjustment instructions to achieve dynamic range adaptive adjustment.

8. The method for adaptively adjusting the dynamic range of an infrared thermal imager according to claim 7, characterized in that: In step 6, the dynamic filtering unit (43) does not filter the independent regions that reach the threshold value of deviation from the noise characteristic.

9. The method for adaptively adjusting the dynamic range of an infrared thermal imager according to claim 7, characterized in that: In step 7, the microcontroller (521) uses a machine learning model to perform in-depth analysis on the input temperature and noise data.

10. The method for adaptively adjusting the dynamic range of an infrared thermal imager according to claim 7, characterized in that: In step 7, the field programmable gate array (522) generates in real time instructions for optimizing the parameter values ​​of the contrast enhancement unit (3) and the dynamic filtering unit (43) based on the gradient descent method and the feedback adjustment mechanism.

Images