Lower limb ischemia assessment system based on thermal imaging key area positioning and temperature difference analysis

Abstract

The invention discloses a lower limb ischemia assessment system based on thermal imaging key area positioning and temperature difference analysis. A human body target is captured and thermal images of different parts are collected through an automatic moving thermal imaging collection unit; carrying out key point detection and region segmentation by utilizing a deep learning model through preprocessing, and accurately dividing an anatomical region; extracting the calibrated area temperature and calculating the temperature difference; finally, lower limb ischemia is evaluated, a result is output, thermal imaging automation and high-precision collection are achieved, and temperature measurement accuracy and standardization are improved. Objective and reliable evaluation is provided, diagnosis efficiency and accuracy are improved, and personal errors are reduced.

Description

Technical Field

[0001] This invention belongs to the field of medical diagnostic technology, specifically relating to a lower limb ischemia assessment system based on thermal imaging key area localization and temperature difference analysis. Background Technology

[0002] With the continuous development and popularization of thermal imaging technology in the field of medical diagnosis, it has demonstrated unique advantages in assisting disease screening and assessment as a non-invasive and radiation-free means of acquiring physiological information. This technology captures the infrared radiation emitted by the surface of an object, converting the invisible temperature distribution into a visible image, thereby reflecting tissue metabolic activity and blood perfusion status. Especially in areas such as circulatory system diseases, inflammatory responses, and early tumor screening, thermal imaging provides valuable auxiliary diagnostic information for clinicians due to its real-time, rapid, and objective characteristics. Precise measurement and in-depth analysis of the surface temperature of specific parts of the human body have become a key technological direction for assessing local physiological function, monitoring pathological changes, and especially blood supply.

[0003] Lower limb ischemia, a common complication of systemic diseases such as atherosclerosis and diabetes, requires early diagnosis and quantitative assessment for guiding clinical treatment and preventing disabling and fatal complications. Traditional diagnostic methods, such as arterial pulsation palpation, Doppler ultrasound, or ankle-brachial index measurement, while each with its own emphasis, generally suffer from limitations such as high subjectivity, complex operation, or inability to provide intuitive information on body surface temperature distribution. Thermal imaging technology, due to its ability to visually present differences in lower limb skin temperature distribution, is considered a potentially effective tool for assessing lower limb blood perfusion status, indirectly indicating ischemia by observing local hypothermic areas.

[0004] However, the application of existing thermal imaging in lower limb ischemia assessment still faces numerous technical challenges. Specifically, current methods lack automation in thermal data acquisition, often requiring manual intervention in camera movement and focusing. Furthermore, they lack intelligent mechanisms for identifying and accurately capturing different key body parts such as feet, hands, face, and legs, resulting in low image acquisition efficiency and inconsistent image quality. In image analysis, for specific anatomical regions of the lower limbs, such as the front and back of the left and right feet, and the inner and outer sides of the left and right legs, current technologies are significantly inadequate in automatically identifying and accurately locating these fixed measurement points for standardized temperature labeling. This makes the quantitative analysis of local temperature differences lack reliable benchmarks and repeatability. Moreover, for auxiliary diagnostic areas such as the face and hands, current methods have failed to achieve automated identification and temperature labeling of multiple fixed locations, such as the sides and front of the left and right faces, and the front and back of the left and right hands, limiting the comprehensiveness of the assessment and the depth of multi-dimensional analysis. The aforementioned shortcomings severely restrict the intelligent, standardized, and clinical application of thermal imaging in lower limb ischemia assessment, and there is an urgent need for an innovative method that can achieve full-process automation, high-precision regional positioning, and multi-site collaborative temperature difference analysis. Summary of the Invention

[0005] The technical problem to be solved by this invention is that the thermal imaging assessment process lacks automation and standardization, and it is difficult to accurately identify key areas of the body and perform temperature difference analysis of multiple parts.

[0006] To address the aforementioned technical problems, the present invention provides a lower limb ischemia assessment system based on thermal imaging key area localization and temperature difference analysis, comprising: The motion control unit receives and responds to the thermal imaging acquisition request, drives the electrically connected thermal imaging acquisition unit to move vertically up and down, and controls the thermal imaging acquisition unit to stop moving when the visual perception unit identifies and captures the first and second specific body parts of the human target. After the thermal imaging unit acquires the first set of thermal images and the second set of thermal images, it transmits them to the motion control unit. The first specific body part includes the feet and legs, and the second specific body part includes the face and hands. The motion control unit drives the electrically connected image processing unit to call the image preprocessing module to preprocess the first set of thermal images and the second set of thermal images, and transmits the preprocessed first set of thermal images and the preprocessed second set of thermal images to the motion control unit. The motion control unit drives the electrically connected identification and segmentation unit to perform key point detection based on the pre-processed first set of thermal images and the pre-processed second set of thermal images, and outputs the two-dimensional key point coordinates of the human skeleton. The key points include the ankle joint, knee joint, hip joint, wrist joint, elbow joint, shoulder joint, nose tip, corner of eye and earlobe. Based on the two-dimensional key point coordinates of the human skeleton, it identifies and segments multiple first specific anatomical regions and multiple second specific anatomical regions. The first specific anatomical region is used to describe the regions after finely dividing the front and back regions of the left and right feet and the inner and outer regions of the left and right legs, with the ankle joint, knee joint and hip joint as reference points. The second specific anatomical region is used to describe the regions after finely dividing the inner region of the left and right face, the front region of the left and right face, the left and right neck region, the front region of the left and right hands and the back region of the left and right hands. The motion control unit drives the temperature extraction and calibration unit, which is electrically connected to the control unit. It selects preset fixed marker points within a first and second specific anatomical region and extracts precise temperature values ​​to obtain a temperature dataset. The preset fixed marker points describe the average temperature points or specific anatomical landmarks within the first and second specific anatomical regions. The specific anatomical landmarks describe the selected landmarks on the front and back of the left and right feet, the inner and outer sides of the left and right legs, the inner and front of the left and right faces, the left and right necks, the front of the left and right hands, and the back of the left and right hands. Based on the temperature dataset, it calls the environmental temperature compensation module and the emissivity calibration module. By acquiring environmental temperature and humidity data from environmental sensors, it compensates for the influence of background radiation and atmospheric attenuation on the temperature measurement results, thereby performing environmental temperature compensation on the extracted temperature values. Based on preset emissivity parameters for different skin regions, it performs emissivity calibration and adjustment. According to the blackbody radiation theory formula, it converts the extracted temperature values ​​into true surface temperature values, achieving environmental temperature compensation and emissivity calibration to obtain standardized temperature data. The motion control unit drives the temperature difference analysis and evaluation unit, which is connected to the control temperature difference. Based on standardized temperature data and according to preset diagnostic rules, it calls the temperature difference calculation module to calculate the temperature difference between the first and second specific anatomical regions. It then calls the medical diagnostic evaluation module to comprehensively consider the temperature difference and the absolute temperature values ​​of each specific anatomical region as preset medical diagnostic thresholds, and performs ischemia assessment model inference to conduct a preliminary assessment of lower limb ischemia. When the temperature difference between the same parts of the left and right lower limbs exceeds the preset threshold, it is marked as asymmetry, indicating a possible risk of lower limb ischemia, and outputs an assessment result including the temperature difference, a warning of potential risks, and suggestions.

[0007] Preferably, the thermal imaging acquisition unit moves vertically up and down from its initial position within a specific vertical scanning range of 0m to 1.8m.

[0008] Preferably, the thermal imaging acquisition unit uses an encoder to enable the servo motor to provide position feedback, and drives a linear slide rail or a multi-axis robotic arm to achieve vertical lifting and lowering movement based on the feedback position.

[0009] Preferably, the thermal imaging unit uses an uncooled microbolometer array sensor with an array size of 384×384 pixels, a temperature measurement range of -20℃ to 150℃, and an accuracy of ±0.2℃.

[0010] Preferably, the preprocessing includes performing two-dimensional Gaussian filtering on the first set of thermal images and the second set of thermal images to remove noise, and performing histogram equalization to enhance image contrast.

[0011] Preferably, the temperature difference between the first specific anatomical region and the second specific anatomical region includes: the temperature difference on the same side of the left and right feet, the temperature difference on the same side of the left and right legs, the overall temperature difference between the left and right feet, the overall temperature difference between the left and right legs, and the longitudinal temperature difference between different segments of the lower limb.

[0012] This invention provides a lower limb ischemia assessment system based on key area localization and temperature difference analysis using thermal imaging. It captures human targets and acquires thermal images of different body parts through an automated mobile thermal imaging acquisition unit. After preprocessing, a deep learning model is used for key point detection and region segmentation to accurately delineate anatomical regions. The system extracts the calibrated region temperature and calculates the temperature difference. Finally, it assesses lower limb ischemia and outputs the results. This system achieves automated and high-precision thermal imaging acquisition, improving temperature measurement accuracy and standardization. It provides an objective and reliable assessment, enhancing diagnostic efficiency and accuracy while reducing human error. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the overall architecture of the lower limb ischemia assessment system based on thermal imaging key area localization and temperature difference analysis proposed in this embodiment of the invention. Figure 2 This is a schematic diagram of the core principle framework for intelligent processing and temperature analysis of key areas of thermal imaging images in this embodiment of the invention. Figure 3 This is a logical flowchart of the automated acquisition of thermal imaging images and human target recognition in an embodiment of the present invention; Figure 4 This is a logical flowchart of the lower limb ischemia assessment method based on thermal imaging key area localization and temperature difference analysis proposed in the embodiments of the present invention. Figure 5 This is a logical flowchart of temperature extraction, calibration, and temperature difference analysis for a specific anatomical region in an embodiment of the present invention. Detailed Implementation

[0014] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0015] like Figure 1 and Figure 2 As shown, this embodiment of the invention provides a lower limb ischemia assessment system based on thermal imaging key area localization and temperature difference analysis. It aims to perform preliminary screening and assessment of blood perfusion in the lower limbs in a non-contact, non-invasive manner, assisting clinical practice in the early identification of lower limb ischemic diseases. The system includes: like Figure 3 and Figure 4 As shown, the motion control unit acquires and responds to the thermal imaging acquisition request, drives the electrically connected thermal imaging acquisition unit to move vertically up and down within a specific vertical scanning range from the initial position. When the visual perception unit identifies and captures the first and second specific body parts of the human target through a multi-sensor fusion algorithm, it controls the thermal imaging acquisition unit to stop moving and controls the thermal imaging unit to acquire the first set of thermal images of the first specific body part and the second set of thermal images of the second specific body part, and then transmits them to the motion control unit.

[0016] The first specific body part includes the feet and legs, and the second specific body part includes the face and hands.

[0017] The thermal imaging acquisition unit uses an encoder to enable the servo motor to provide position feedback, and drives a linear slide rail or multi-axis robotic arm to perform vertical lifting and lowering movements within a specific vertical scanning range based on the feedback position.

[0018] The specific vertical scanning range is from 0m to 1.8m.

[0019] The thermal imaging unit uses an uncooled microbolometer array sensor with an array size of 384×384 pixels, a temperature measurement range of -20℃ to 150℃, and an accuracy of ±0.2℃. The optical system of the thermal imaging unit uses a germanium lens with a field of view of 45 degrees.

[0020] The visual perception unit includes a visible light camera and a depth sensor. The visual perception unit tracks the key points of the human skeleton in real time to assist the thermal imaging acquisition unit in positioning and motion control.

[0021] The motion control unit drives the electrically connected image processing unit to call the image preprocessing module to preprocess the first set of thermal images and the second set of thermal images, and then transmits the preprocessed first set of thermal images and the preprocessed second set of thermal images to the motion control unit.

[0022] The preprocessing includes performing two-dimensional Gaussian filtering on the first set of thermal images and the second set of thermal images to remove noise, and performing histogram equalization to enhance image contrast.

[0023] The kernel size of the two-dimensional Gaussian filter is 3×3, and the standard deviation is 0.8.

[0024] like Figure 5As shown, the motion control unit drives the electrically connected recognition and segmentation unit. Based on the pre-processed first set of thermal images and the pre-processed second set of thermal images, it uses a deep learning human pose recognition model to detect key points and outputs the two-dimensional key point coordinates of the human skeleton. The key points include the ankle joint, knee joint, hip joint, wrist joint, elbow joint, shoulder joint, nose tip, corner of the eye, and earlobe. Based on the two-dimensional key point coordinates of the human skeleton, it uses a region segmentation algorithm, morphological operations, and region growing algorithms to accurately identify and segment multiple first specific anatomical regions of a first specific body part and multiple second specific anatomical regions of a second specific body part. The first specific anatomical region is used to describe the region after finely dividing the front and back regions of the left and right feet, as well as the inner and outer regions of the left and right legs, using the ankle joint, knee joint, and hip joint as reference points and using predefined geometric templates or semantic segmentation networks. The second specific anatomical region is used to describe the region after finely dividing the inner region of the left and right face, the front region of the left and right face, the left and right neck regions, the front region of the left and right hands, and the back region of the left and right hands.

[0025] The deep learning human pose recognition model is based on a pose estimation algorithm using a convolutional neural network, and is trained on a dataset of historical human thermal imaging images and corresponding key point annotations.

[0026] The region segmentation algorithm is obtained by training the pose recognition model.

[0027] The motion control unit drives the temperature extraction and calibration unit, which is electrically connected to the control unit. Within each segmented first and second specific anatomical region, preset fixed marker points are selected and precise temperature values ​​are extracted to obtain a temperature dataset. The preset fixed marker points are used to describe the average temperature points or specific anatomical landmarks within the first and second specific anatomical regions. The specific anatomical landmarks are used to describe the selected landmarks on the front and back areas of the left and right feet, the inner and outer areas of the left and right legs, the inner and front areas of the left and right faces, the left and right neck areas, the front and back areas of the left and right hands, and the back areas of the left and right hands. Based on the temperature dataset, the unit calls the environmental temperature compensation module and the emissivity calibration module. By acquiring environmental temperature and humidity data from environmental sensors, the unit uses a correction algorithm to compensate for the influence of background radiation and atmospheric attenuation on the temperature measurement results, thereby performing environmental temperature compensation on the extracted temperature values. Based on preset emissivity parameters for different skin regions, the unit performs emissivity calibration and adjustment. According to the blackbody radiation theory formula, the extracted temperature values ​​are converted into true surface temperature values, realizing environmental temperature compensation and emissivity calibration, and obtaining standardized temperature data.

[0028] The motion control unit drives the temperature difference analysis and evaluation unit, which is connected to the control temperature difference. Based on standardized temperature data and according to preset diagnostic rules, it calls the temperature difference calculation module to calculate the temperature difference between the first and second specific anatomical regions. Then, it calls the medical diagnostic evaluation module to comprehensively consider the temperature difference and the absolute temperature values ​​of each specific anatomical region as preset medical diagnostic thresholds through a multi-parameter fusion algorithm. This performs ischemia assessment model inference and makes a preliminary assessment of the lower limb ischemia. When the temperature difference between the same parts of the left and right lower limbs exceeds the preset threshold, it is marked as asymmetry, indicating a possible risk of lower limb ischemia. The system then outputs an evaluation result that includes the temperature difference, a warning of the possible risk, and suggestions.

[0029] The preset threshold is 0.5℃ to 1.0℃.

[0030] The temperature difference between the first specific anatomical region and the second specific anatomical region includes: the temperature difference on the same side of the left and right feet, the temperature difference on the same side of the left and right legs, the overall temperature difference between the left and right feet, the overall temperature difference between the left and right legs, and the longitudinal temperature difference between different segments of the lower limb.

[0031] The formula for calculating the temperature difference is: Temperature difference = Temperature of the first part - Temperature of the second part The multi-parameter fusion algorithm is trained based on the reasoning logic of expert rules and the predictive ability of machine learning models.

[0032] The lower limb ischemia assessment system based on thermal imaging key area localization and temperature difference analysis provided in this invention has achieved automated acquisition of thermal imaging data, accurate identification and temperature marking of key areas in multiple parts of the human body, and temperature difference analysis based on anatomical and physiological principles through an integrated intelligent sensing and control mechanism, thereby improving the efficiency and accuracy of lower limb ischemia and other related physiological indicators assessment.

[0033] Compared with the prior art, the advantages and positive effects of the embodiments of the present invention are as follows: The embodiments of the present invention achieve automated and high-precision acquisition of thermal imaging images through an integrated intelligent thermal imaging acquisition unit and motion control unit, avoiding the problems of positional deviation and poor data consistency caused by traditional manual operation, and improving acquisition efficiency.

[0034] The combination of visual perception unit and deep learning human pose recognition model enables the system to accurately identify and locate multiple key areas of the human body, including the lower limbs, face and hands, and automatically select preset fixed marker points in these areas for temperature extraction, which greatly improves the accuracy and standardization of temperature measurement points.

[0035] The embodiments of the present invention further utilize a temperature difference analysis and evaluation unit to calculate temperature difference values ​​based on standardized multi-site temperature data and in accordance with strict medical diagnostic rules, and perform quantitative evaluation. This provides objective and reliable lower limb ischemia risk assessment results, improving the scientific nature and efficiency of diagnosis.

[0036] The embodiments of the present invention reduce reliance on operator experience and human error through automated processes, and provide non-contact, efficient and highly repeatable data support for medical diagnosis.

[0037] This invention provides a system-level solution that deeply integrates thermal imaging technology with advanced machine vision, deep learning, and precision motion control technologies, forming an innovative paradigm with non-obviousness and overcoming the shortcomings of existing technologies in terms of automation, standardization, and refined analysis.

Claims

1. A system for lower extremity ischemia assessment based on thermographic key area positioning and temperature difference analysis, characterized in that, The application relates to a thermal imaging system and a method for acquiring and processing thermal images. The motion control unit acquires a thermal imaging acquisition request and responds, drives the electrically connected thermal imaging acquisition unit to vertically move, controls the thermal imaging acquisition unit to stop moving when the visual perception unit identifies and captures a first specific body part and a second specific body part in a human body target, and controls the thermal imaging unit to acquire a first group of thermal images and a second group of thermal images and then transmit the first group of thermal images and the second group of thermal images to the motion control unit, wherein the first specific part includes feet and legs, and the second specific part includes a face and hands. The motion control unit drives the electrically connected image processing unit to call an image preprocessing module to pre-process the first group of thermal images and the second group of thermal images, and then transmit the pre-processed first group of thermal images and the pre-processed second group of thermal images to the motion control unit. The motion control unit drives the electrically connected identification and segmentation unit to perform key point detection according to the pre-processed first group of thermal images and the pre-processed second group of thermal images, output two-dimensional key point coordinates of a human skeleton, and identify and segment a plurality of first specific anatomical regions and a plurality of second specific anatomical regions according to the two-dimensional key point coordinates of the human skeleton, wherein the first specific anatomical region is used for describing regions obtained by finely dividing front and back regions of left and right feet and inner and outer regions of left and right legs with the ankle joint, the knee joint and the hip joint as reference points, and the second specific anatomical region is used for describing regions obtained by finely dividing inner regions of left and right faces, front regions of left and right faces, left and right neck regions, front regions of left and right hands and back regions of left and right hands. The motion control unit drives the electrically connected control temperature extraction and calibration unit to select preset fixed mark points in the first specific anatomical region and the second specific anatomical region and extract accurate temperature values to obtain a temperature data set, the preset fixed mark points are used for describing uniform points or specific anatomical mark points in the first specific anatomical region and the second specific anatomical region, the specific anatomical mark points are used for describing mark points selected from front regions of left and right feet, back regions of left and right feet, inner regions of left and right legs, outer regions of left and right legs, inner regions of left and right faces, front regions of left and right faces, left and right neck regions, front regions of left and right hands and back regions of left and right hands, the temperature data set is used for calling an environment temperature compensation module and a radiation rate calibration module, acquiring environment temperature and humidity data collected by an environment sensor, compensating influences of background radiation and atmospheric attenuation on temperature measurement results, performing environment temperature compensation on the extracted temperature values, performing radiation rate calibration adjustment based on preset radiation rate parameters of different skin regions, converting the extracted temperature values into real surface temperature values according to a blackbody radiation theory formula, realizing environment temperature compensation and radiation rate calibration, and obtaining standardized temperature data. The motion control unit drives the electrically connected control temperature difference analysis and evaluation unit, based on standardized temperature data, according to preset diagnostic rules, calls a temperature difference calculation module, calculates the temperature difference value between the first specific anatomical region and the second specific anatomical region, calls a medical diagnosis evaluation module to comprehensively consider the temperature difference value and the absolute temperature value of each specific anatomical region as a preset medical diagnosis threshold, performs ischemia evaluation model reasoning, and preliminarily evaluates the lower limb ischemia condition. When the temperature difference value of the homologous parts of the left and right lower limbs exceeds the preset threshold, it is marked as having asymmetry, suggesting that there may be a risk of lower limb ischemia, and an evaluation result containing the temperature difference value, a suggestion that there may be a risk, and a suggestion is output.

2. The lower extremity ischemia assessment system based on thermographic key area positioning and temperature difference analysis according to claim 1, characterized in that, The thermal imaging acquisition unit moves vertically within a specific vertical scanning range of 0 m to 1.8 m from an initial position.

3. The lower extremity ischemia assessment system based on thermographic key area positioning and temperature difference analysis of claim 1, wherein, The thermal imaging acquisition unit feeds back the position of the servo motor through the encoder, and drives the linear slide or multi-axis mechanical arm to realize vertical lifting movement according to the feedback position.

4. The lower extremity ischemia assessment system based on thermographic key area positioning and temperature difference analysis of claim 1, wherein, The thermal imaging unit selects a non-cooled microbolometer array sensor, which has an array size of 384×384 pixels, a temperature measurement range of -20°C to 150°C, and an accuracy of ±0.2°C.

5. The thermal imaging based key area positioning and temperature difference analysis based lower extremity ischemia assessment system of claim 1, wherein, The preprocessing includes two-dimensional Gaussian filtering of the first and second groups of thermal images to filter out noise, and histogram equalization to enhance image contrast.

6. The thermal imaging based key area positioning and temperature difference analysis based lower extremity ischemia assessment system of claim 1, wherein, The temperature difference value between the first specific anatomical region and the second specific anatomical region includes the temperature difference of the same side of the left and right feet, the temperature difference of the same side of the left and right legs, the overall temperature difference between the left and right feet, the overall temperature difference between the left and right legs, and the longitudinal temperature difference between different segments of the lower limbs.

Images