Stroke monitoring cervical pillow system based on infrared thermal imaging and early warning method

Abstract

The application discloses a stroke monitoring neck pillow system based on infrared thermal imaging, which comprises a medical memory cotton pillow body, an ultrathin polyester fiber fabric pillowcase, a silica gel sleeve and a stroke detection integrated device. The stroke detection integrated device comprises an infrared thermal imager, temperature sensors, a pressure sensor and a control panel. The infrared thermal imager, the temperature sensors and the pressure sensor are connected with the control panel. With the vertical center line of the pillow body top view as a reference, the control panel, the pressure sensor, the temperature sensors and the infrared thermal imager are arranged from top to bottom in sequence. The control panel is arranged in the upper cavity in the pillow body. The pressure sensor is arranged in the middle cavity in the pillow body. Two temperature sensors are arranged on the surface of the pillow body and are symmetrically arranged with respect to the vertical center line. Two infrared thermal imagers are arranged on the surface of the pillow body and are symmetrically and obliquely arranged with respect to the vertical center line. The silica gel sleeve wraps the pressure sensor and the control panel. The pillowcase wraps the infrared thermal imager, the temperature sensors and the pillow body.

Description

Technical Field

[0001] This invention relates to the field of stroke monitoring and early warning technology, and more specifically, to a stroke monitoring neck pillow system and early warning method based on infrared thermal imaging suitable for home settings. Background Technology

[0002] Current stroke early warning technologies designed for home use suffer from the following core shortcomings, severely hindering their widespread application. Traditional early warning systems rely on specialized equipment such as MRI and ultrasound imaging, making seamless home monitoring impossible. While existing home devices are equipped with multimodal sensors, the detection process requires users to actively press buttons and complete specified actions (such as fine motor skills tests), posing operational barriers for the elderly and patients experiencing sudden symptoms.

[0003] Wearable devices (such as smart bracelets) can easily cause skin allergies with prolonged wear, and nighttime monitoring may disrupt sleep. Furthermore, mattress-based solutions, limited by size and firmness, struggle to balance sleep comfort with the accuracy of physiological signal acquisition. Studies show that approximately 68% of users discontinue use for more than a month due to comfort issues. Complex factors in the home environment, such as temperature fluctuations (±5℃) and electromagnetic interference (from WiFi / Bluetooth devices), reduce the stability of infrared thermal image feature extraction. Summary of the Invention

[0004] The purpose of this invention is to overcome the defects and deficiencies of the prior art and provide a stroke monitoring neck pillow system and early warning method based on infrared thermal imaging. This system enables stroke detection to be routine and made possible at home, allowing users to understand their own stroke risk information in advance. Users can then go to the hospital for professional examination and treatment based on the risk warning information.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A stroke monitoring neck pillow system based on infrared thermal imaging includes a medical memory foam pillow body, an ultra-thin polyester fiber fabric pillowcase, a silicone sleeve, and a stroke detection integrated device. The stroke detection integrated device includes an infrared thermal imager, a temperature sensor, a pressure sensor, and a control board. The infrared thermal imager, temperature sensor, and pressure sensor are respectively connected to the control board.

[0007] With the vertical centerline of the pillow body as a reference, the control board, pressure sensor, temperature sensor, and infrared thermal imager are arranged sequentially from top to bottom. The control board is located in the upper cavity inside the pillow body, the pressure sensor is located in the middle cavity inside the pillow body, the two temperature sensors are respectively located on the surface of the pillow body and are symmetrically arranged with respect to the vertical centerline of the pillow body, and the two infrared thermal imagers are respectively located on the surface of the pillow body and are symmetrically inclined with respect to the vertical centerline of the pillow body. The silicone sleeves cover the pressure sensor and the control board, and the pillowcase covers the infrared thermal imager, the temperature sensor, and the pillow body.

[0008] The infrared thermal imager is used to acquire infrared thermal images of both sides of the spine during sleep. The temperature sensor is used to acquire precise temperature data of both sides of the spine during sleep. The pressure sensor is used to acquire pressure data of the pillow during sleep. The control board is used to process and analyze the infrared thermal images, extract the detected temperature data of both sides of the spine, and combine it with the precise temperature data to perform temperature analysis and judgment to obtain the temperature difference ΔT between the two sides of the spine.

[0009] The control board is used to calculate the longest PDT (postural ischemia-reperfusion index) of the user in the same sleeping position based on pressure data, and to calculate the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) by combining the temperature difference ΔT between the two sides of the spine during the user's sleep. The control board is also used to determine the stroke risk level based on the comprehensive risk index (SRI) of the user during sleep, generate stroke monitoring and early warning analysis results, and send the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display.

[0010] Furthermore, the control board integrates a main control chip, a voltage conversion unit, a download circuit, a charging circuit, a Bluetooth circuit, and a timer circuit. An external power supply is connected to the voltage conversion unit or the charging circuit, the main control chip is connected to the voltage conversion unit, and the main control chip and the voltage conversion unit are respectively connected to the download circuit, the Bluetooth circuit, and the timer circuit.

[0011] Furthermore, the voltage conversion unit includes a 3.7V to 5V circuit, a 5V to 3.3V circuit, and a 5V to -5V circuit. The 3.7V to 5V circuit and the 5V to -5V circuit provide power to the infrared thermal imager and the pressure sensor, while the 5V to 3.3V circuit provides power to the main control chip, the download circuit, the Bluetooth circuit, and the temperature sensor.

[0012] Furthermore, the external power supply input is 3.7V, and the main control chip is an STM32F103C8T6 chip; the download circuit is connected to the main control chip through the CLK interface and DIO interface, and uses an ST-LINK programmer with Keil5 software for programming; the charging circuit is connected to the external power supply 3.7V through a sliding switch, and uses a Type-C USB interface; the Bluetooth circuit is connected to the main control chip through the USART interface, and uses the Bluetooth specification BLEV5.3+BR+EDR.

[0013] Furthermore, the pressure sensor includes four strain gauges arranged in an array. The strain gauge wire grid size is 3×3mm, the strain gauge base size is 6.8×4.2mm, the strain gauge sensitivity coefficient is 2.0±1%, the strain gauge output voltage is -3V to +4V, and the strain gauge output voltage accuracy is 0.001V.

[0014] A stroke monitoring and early warning method based on infrared thermal imaging, applied to any of the above-mentioned stroke monitoring neck pillow systems based on infrared thermal imaging, includes the following steps:

[0015] S1: The infrared thermal imager collects infrared thermal images of both sides of the user's spine during sleep, the temperature sensor collects precise temperature data of both sides of the user's spine during sleep, and the pressure sensor collects pressure data of the user on the pillow during sleep.

[0016] S2: The control board processes and analyzes the infrared thermal imaging image, extracts the detected temperature data on both sides of the spine, and combines the accurate temperature data to perform temperature analysis and judgment on the detected temperature data to obtain the temperature difference ΔT on both sides of the spine.

[0017] S3: The control board calculates the longest PDT (postural ischemia index) of the user in the same sleeping position based on the pressure data, and calculates the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) by combining the temperature difference ΔT between the two sides of the spine during the user's sleep.

[0018] S4: The control board determines the stroke risk level based on the user's Comprehensive Risk Index (SRI) during sleep, generates stroke monitoring and early warning analysis results, and sends the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display.

[0019] Further, step S2 includes:

[0020] The infrared thermal imaging image is preprocessed by using Gaussian filtering to remove background noise and sensing errors. The thermal data in the infrared thermal imaging image is converted into temperature values, and the temperature range of a normal human body is selected to distinguish human body temperature from ambient temperature. The infrared thermal imaging image is segmented to separate the human body area from the environment area, and Canny edge detection is used to extract the human body contour from the infrared thermal imaging image.

[0021] The position of the spine in the infrared thermal image is used as a reference line for division. The left and right sides are divided according to the relative positions of the brain and spine. The center line is inferred by the temperature change of the spine region in the infrared thermal image, and the position of the center line is determined by an algorithm based on temperature gradient. The temperature change of each column of the infrared thermal image is calculated to find the place with the most obvious temperature change. The infrared thermal image is divided into left and right sides according to the center line.

[0022] Furthermore, step S2 also includes:

[0023] After dividing the area into left and right sections, extract the temperature data from the left and right sections respectively, and calculate the temperature distribution of the corresponding areas; calculate the difference in average temperature between the two sections. Determine if there is a large temperature difference between the two sides; if the average temperature difference is significant. This indicates a significant temperature difference between the two sides, requiring temperature difference calculation; the difference between the lowest temperatures on both sides of the spine is calculated, i.e., |T 左 -T 右 |=ΔT;

[0024] Using precise temperature data provided by the temperature sensor, the system assists in determining the temperature regions on both sides; it checks whether any low-temperature regions appear in the temperature sensor's measurement area; if the temperature sensor's value differs significantly from the infrared thermal imaging result, and the low-temperature region in the infrared image does not appear in the temperature sensor's area, the temperature sensor's value helps determine the exact location of the low-temperature region; if the low-temperature region is within the temperature sensor's measurement area, the temperature sensor's value is used to correct the temperature value in the infrared thermal imaging image to improve the accuracy of the temperature difference calculation; and it outputs the temperature difference ΔT between the two sides of the spine.

[0025] Furthermore, the control board calculates the longest PDT (Prolonged Sleep Time) for the user in the same sleeping position based on the pressure data, specifically:

[0026] The pressure value of the four strain gauges is set to P. A P B P C and P D Before analyzing the turning-over behavior, all stress data were normalized as follows:

[0027]

[0028] In the formula, P i P represents the pressure value of strain gauge i; i ′ The normalized pressure value of strain gauge i; P max P represents the maximum pressure value of the strain gauge. min This is the minimum pressure value of the strain gauge;

[0029] When monitoring whether a user turns over, the user's sleep behavior is divided into static state and dynamic state. The static state indicates that the user has not turned over, and the dynamic state indicates that the user has turned over.

[0030] By calculating the rate of change of the pressure values ​​of the four strain gauges, it can be determined whether the user is currently in a dynamic state.

[0031] The formula for calculating the rate of change of pressure is: ΔP i =P i ′ (t)-P i ′ (t-1);

[0032] In the formula, ΔP i P represents the rate of change of pressure value at strain gauge i. i ′ (t) represents the pressure value of strain gauge i at time t after normalization; P i ′ (t-1) represents the pressure value of strain gauge i at time t-1 after normalization.

[0033] Calculate the pressure changes of all strain gauges, and then calculate the sum of the pressure changes of all strain gauges. This sum will be used as an indicator of the user's recovery. The calculation formula is as follows:

[0034] In the formula, ΔP total The sum of the rates of change of pressure values ​​of all strain gauges; if ΔP total If the value exceeds the preset threshold, the user is considered to have turned things around.

[0035] The timing starts when pressure occurs, stops each time the user turns over, and records the duration. A new timing begins, with the subsequent duration overwriting the previous one, but not stored. The timing ends when the pressure reaches 0, yielding the longest physical time (PDT) for the user in the same sleeping position for each time period.

[0036] Furthermore, the positional ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) are calculated as follows:

[0037] The formula for calculating the postural ischemia index (PII) is as follows:

[0038]

[0039] The formula for calculating the Narrow Temperature Index (STI) is:

[0040] The formula for calculating the Comprehensive Risk Index (SRI) is: SRI = 0.6 × min(STI, 150) + 0.4 × PII.

[0041] Compared with existing technologies, this invention uses an infrared thermal imager to acquire infrared thermal images of both sides of the user's spine during sleep, a temperature sensor to collect precise temperature data of both sides of the spine during sleep, and a pressure sensor to collect pressure data of the user's pillow during sleep. The control board calculates the Stroke Monitoring and Early Warning Analysis (SRI) using a comprehensive risk index and sends the results to the mobile user terminal. This application enables routine and home-based stroke detection, allowing users to understand their own stroke risk information in advance, and enabling users to go to the hospital for professional examination and treatment based on the risk warning information. Attached Figure Description

[0042] Figure 1 This is a top view schematic diagram of the occipital body after fluoroscopy and the integrated device for stroke detection.

[0043] Figure 2 This is a side view schematic diagram of the occipital body after fluoroscopy and the integrated device for stroke detection.

[0044] Figure 3 This is a schematic diagram showing the connection between the integrated stroke detection device and the mobile user terminal.

[0045] Figure 4 This is a flowchart illustrating a stroke monitoring and early warning method based on infrared thermal imaging.

[0046] Explanation of icon numbers:

[0047] 1-Infrared thermal imager; 2-Temperature sensor; 3-Pressure sensor; 4-Control board. Detailed Implementation

[0048] The following description, in conjunction with the accompanying drawings and specific embodiments, further illustrates the present invention: a stroke monitoring neck pillow system and early warning method based on infrared thermal imaging.

[0049] Please see Figure 1 and Figure 2This invention discloses a stroke monitoring neck pillow system based on infrared thermal imaging, including a medical memory foam pillow body, an ultra-thin polyester fiber fabric pillowcase, a silicone sleeve, and a stroke detection integrated device. The stroke detection integrated device includes an infrared thermal imager 1, a temperature sensor 2, a pressure sensor 3, and a control board 4. The infrared thermal imager 1, the temperature sensor 2, and the pressure sensor 3 are respectively connected to the control board 4.

[0050] With the vertical centerline of the pillow body as a reference, the control board 4, pressure sensor 3, temperature sensor 2, and infrared thermal imager 1 are arranged from top to bottom. The control board 4 is located in the upper cavity inside the pillow body, the pressure sensor 3 is located in the middle cavity inside the pillow body, the two temperature sensors 2 are respectively located on the surface of the pillow body and are symmetrically arranged with respect to the vertical centerline of the pillow body, and the two infrared thermal imagers 1 are respectively located on the surface of the pillow body and are symmetrically inclined with respect to the vertical centerline of the pillow body. The silicone sleeves cover the pressure sensor 3 and the control board 4 respectively, and the pillowcases cover the infrared thermal imager 1, the temperature sensor 2, and the pillow body.

[0051] Infrared thermal imager 1 is used to acquire infrared thermal images of both sides of the spine during the user's sleep; temperature sensor 2 is used to acquire precise temperature data of both sides of the spine during the user's sleep; pressure sensor 3 is used to acquire pressure data of the pillow body during the user's sleep; control board 4 is used to process and analyze the infrared thermal images, extract the detected temperature data of both sides of the spine, combine the precise temperature data, perform temperature analysis and judgment on the detected temperature data, and obtain the temperature difference ΔT between both sides of the spine.

[0052] Control board 4 is used to calculate the longest PDT (postural ischemia index) of the user in the same sleeping position based on pressure data, and to calculate the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) by combining the temperature difference ΔT between the two sides of the spine during the user's sleep. Control board 4 is used to determine the stroke risk level based on the comprehensive risk index (SRI) of the user during sleep, generate stroke monitoring and early warning analysis results, and send the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display.

[0053] The control board 4 integrates a main control chip, a voltage conversion unit, a download circuit, a charging circuit, a Bluetooth circuit, and a timer circuit. An external power supply is connected to the voltage conversion unit or the charging circuit. The main control chip is connected to the voltage conversion unit. The main control chip and the voltage conversion unit are respectively connected to the download circuit, the Bluetooth circuit, and the timer circuit.

[0054] The voltage conversion unit includes a 3.7V to 5V circuit, a 5V to 3.3V circuit, and a 5V to -5V circuit. The 3.7V to 5V circuit and the 5V to -5V circuit provide power to the infrared thermal imager 1 and the pressure sensor 3, while the 5V to 3.3V circuit provides power to the main control chip, the download circuit, the Bluetooth circuit, and the temperature sensor 2.

[0055] The external power input is 3.7V, and the main control chip is an STM32F103C8T6. The download circuit connects to the main control chip via the CLK and DIO interfaces, and uses an ST-LINK programmer with Keil 5 software for programming. The charging circuit connects to the external 3.7V power supply via a slide switch and uses a Type-C USB interface. The Bluetooth circuit connects to the main control chip via a USART interface, using the Bluetooth specification BLEV5.3+BR+EDR. The timer circuit acts as a timer for the main control chip, calculating the user's usage time of the neck pillow.

[0056] The pressure sensor 3 includes four strain gauges arranged in an array. The strain gauge wire grid size is 3×3mm, the strain gauge base size is 6.8×4.2mm, the strain gauge sensitivity coefficient is 2.0±1%, the strain gauge output voltage is -3V to +4V, and the strain gauge output voltage accuracy is 0.001V.

[0057] Please see Figure 3 and Figure 4 The present invention also discloses a stroke monitoring and early warning method based on infrared thermal imaging, applied to the stroke monitoring neck pillow system based on infrared thermal imaging described in any of the above-mentioned embodiments, comprising the following steps:

[0058] S1: The infrared thermal imager collects infrared thermal images of both sides of the user's spine during sleep, the temperature sensor collects precise temperature data of both sides of the user's spine during sleep, and the pressure sensor collects pressure data of the user on the pillow during sleep.

[0059] S2: The control board processes and analyzes the infrared thermal imaging image, extracts the detected temperature data on both sides of the spine, and combines the accurate temperature data to perform temperature analysis and judgment on the detected temperature data to obtain the temperature difference ΔT on both sides of the spine.

[0060] S3: The control board calculates the longest PDT (postural ischemia index) of the user in the same sleeping position based on the pressure data, and calculates the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) by combining the temperature difference ΔT between the two sides of the spine during the user's sleep.

[0061] S4: The control board determines the stroke risk level based on the user's Comprehensive Risk Index (SRI) during sleep, generates stroke monitoring and early warning analysis results, and sends the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display.

[0062] Step S2: The control board processes and analyzes the infrared thermal imaging image, extracts the detected temperature data on both sides of the spine, and combines it with precise temperature data to perform temperature analysis and judgment, obtaining the temperature difference ΔT between the two sides of the spine, including:

[0063] The infrared thermal imaging image is preprocessed by using Gaussian filtering to remove background noise and sensing errors. The thermal data in the infrared thermal imaging image is converted into temperature values, and the temperature range of a normal human body is selected to distinguish human body temperature from ambient temperature. The infrared thermal imaging image is segmented to separate the human body area from the environment area, and Canny edge detection is used to extract the human body contour from the infrared thermal imaging image.

[0064] The position of the spine in the infrared thermal image is used as a reference line for division. The left and right sides are divided according to the relative positions of the brain and spine. The center line is inferred by the temperature change of the spine region in the infrared thermal image, and the position of the center line is determined by an algorithm based on temperature gradient. The temperature change of each column of the infrared thermal image is calculated to find the place with the most obvious temperature change. The infrared thermal image is divided into left and right sides according to the center line.

[0065] In step S2, the control board processes and analyzes the infrared thermal imaging image, extracts the detected temperature data on both sides of the spine, combines it with precise temperature data, performs temperature analysis and judgment on the detected temperature data, and obtains the temperature difference ΔT between the two sides of the spine. This also includes:

[0066] After dividing the area into left and right sections, extract the temperature data from the left and right sections respectively, and calculate the temperature distribution of the corresponding areas; calculate the difference in average temperature between the two sections. Determine if there is a large temperature difference between the two sides; if the average temperature difference is significant. This indicates a significant temperature difference between the two sides, requiring temperature difference calculation; the difference between the lowest temperatures on both sides of the spine is calculated, i.e., |T 左 -T 右 |=ΔT;

[0067] Using precise temperature data provided by the temperature sensor, the system assists in determining the temperature regions on both sides; it checks whether any low-temperature regions appear in the temperature sensor's measurement area; if the temperature sensor's value differs significantly from the infrared thermal imaging result, and the low-temperature region in the infrared image does not appear in the temperature sensor's area, the temperature sensor's value helps determine the exact location of the low-temperature region; if the low-temperature region is within the temperature sensor's measurement area, the temperature sensor's value is used to correct the temperature value in the infrared thermal imaging image to improve the accuracy of the temperature difference calculation; and it outputs the temperature difference ΔT between the two sides of the spine.

[0068] In step S3, the control board calculates the longest PDT (Prolonged Sleep Time) for the user in the same sleeping position based on the pressure data, specifically:

[0069] The pressure value of the four strain gauges is set to P. A P B P C and P D Before analyzing the turning-over behavior, all stress data were normalized as follows:

[0070]

[0071] In the formula, P i P represents the pressure value of strain gauge i; i ′ The normalized pressure value of strain gauge i; P max P represents the maximum pressure value of the strain gauge. min This is the minimum pressure value of the strain gauge;

[0072] When monitoring whether a user turns over, the user's sleep behavior is divided into static state and dynamic state. The static state indicates that the user has not turned over, and the dynamic state indicates that the user has turned over.

[0073] By calculating the rate of change of the pressure values ​​of the four strain gauges, it can be determined whether the user is currently in a dynamic state.

[0074] The formula for calculating the rate of change of pressure is: ΔP i =P i ′ (t)-P i ′ (t-1);

[0075] In the formula, ΔP i P represents the rate of change of pressure value at strain gauge i. i ′ (t) represents the pressure value of strain gauge i at time t after normalization; P i ′ (t-1) represents the pressure value of strain gauge i at time t-1 after normalization.

[0076] Calculate the pressure changes of all strain gauges, and then calculate the sum of the pressure changes of all strain gauges. This sum will be used as an indicator of the user's recovery. The calculation formula is as follows:

[0077] In the formula, ΔP total The sum of the rates of change of pressure values ​​of all strain gauges; if ΔP total If the value exceeds the preset threshold, the user is considered to have turned things around.

[0078] The timer circuit on the control board starts timing when pressure is detected, stops timing each time the user turns over, and records the timing duration. A new timer is started, and the subsequent timing duration overwrites the previous timing duration; it is not stored. Timing ends when the pressure reaches 0, obtaining the longest physical time (PDT) of the user sleeping in the same position for each time period.

[0079] In step S3, the positional ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) are calculated, specifically as follows:

[0080] The formula for calculating the postural ischemia index (PII) is as follows:

[0081]

[0082] The formula for calculating the Narrow Temperature Index (STI) is:

[0083] The formula for calculating the Comprehensive Risk Index (SRI) is: SRI = 0.6 × min(STI, 150) + 0.4 × PII.

[0084] In step S4, after obtaining the SRI (Stroke Risk Index) calculation result, the SRI is used as a parameter for the mean squared error algorithm. A second calculation is performed to obtain the processing result. This processing result is compared with a preset value to determine the user's stroke risk level. If the processing result is less than or equal to preset interval 1, the user's stroke risk level is determined to be low; if the processing result is within preset interval 2, the user's stroke risk level is determined to be medium; and if the processing result is greater than or equal to preset interval 3, the user's stroke risk level is determined to be high. Based on the above determination, the main control chip generates a stroke monitoring and early warning analysis result and sends it to the mobile user terminal via Bluetooth.

[0085] The mobile user terminal processes the stroke monitoring and early warning analysis results and generates different feedback to the user based on the results. For example, if the user's stroke risk level is low, the mobile user terminal pushes a health alert. If the user's stroke risk level is medium, the mobile user terminal indicates that the user needs to seek professional medical advice for prevention and diagnosis and sends a text message reminder to the user. If the user's stroke risk level is high, the mobile user terminal page turns red as a warning, generates a high-risk stroke early warning page, and enables phone permissions to send advanced stroke warning text messages to emergency contacts and family members.

[0086] To ensure the security of user data during transmission between the neck pillow system and the mobile user terminal, this invention employs the Galilean Domain Multiplication (GCM) mode of the Advanced Encryption Standard (AES), specifically the AES-256-GCM encryption algorithm. The specific implementation is as follows: A 256-bit symmetric key is generated dynamically during each session and shared between the communicating parties through a secure key exchange mechanism (such as Diffie-Hellman).

[0087] The core parameters of this algorithm are as follows: Key Size = 256 bits. AES-256-GCM uses a 256-bit key, which makes it highly resistant to brute-force attacks. The Initialization Vector (IV) ensures the uniqueness of each encryption operation, preventing the same plaintext from generating the same ciphertext. The IV is typically 12 bytes (96 bits) long, which is the standard length for GCM mode. The Message Authentication Code (MAC) is used to verify the integrity and authenticity of the data. The MAC is 128 bits long. Additional Authentication Data (AAD) is data that does not need to be encrypted but requires authentication. It is used to generate authentication tags during encryption but is not decrypted during decryption.

[0088] The encryption process is as follows: A 256-bit AES key is generated using a key generator. A 12-byte IV is generated using a secure random number generator. The data is encrypted using the AES-256-GCM algorithm, with the following formula: C = AES-256-GCM(P, K, IV), where C is the encrypted ciphertext, P is the plaintext data, K is the 256-bit key, and IV is the initialization vector used to ensure the uniqueness of each encryption.

[0089] The receiver uses the same key and initialization vector to decrypt the ciphertext and verifies the MAC to ensure that the data has not been tampered with. The decryption process is as follows: the key and IV used during encryption are loaded and decrypted using an algorithm. Decryption is performed using the following formula, including: P = AES - 256 - GCM - DEC(C, K, IV), where P is the decrypted plaintext data.

[0090] The mobile user terminal data storage method is as follows: The neck pillow system user data is stored in a time-series database, using the following data model, including: DataPoint={Time,SensorID,UserID,Value}, where Time is the timestamp, SensorID is the unique identifier of the main control chip, UserID is the unique identifier of the user, and Value is the analysis result of the main control chip.

[0091] The following detailed calculation examples illustrate the stroke monitoring and early warning method based on infrared thermal imaging of the present invention.

[0092] When the user continuously maintains the same sleeping posture for a posture duration time PDT of 120 minutes, and the minimum temperature difference ΔT between the two sides detected by infrared thermal imaging of the spinal region is 1.0°C, the stroke risk grading calculation is performed through the following steps:

[0093] Calculation of Postural Ischemia Index PII: According to the piecewise function formula, since PDT = 120 minutes (satisfying the condition of 90 < PDT ≤ 180 minutes), the result is obtained by substitution calculation

[0094]

[0095] Calculation of Stenosis Temperature Difference Index STI: Based on the temperature difference ΔT = 1.0°C, according to the formula

[0096] STI=50·ln(1.0+1)+20·1.0=50·0.693+20·1.0=34.65+20=54.65

[0097] Calculation of Synthetic Risk Index SRI:

[0098] SRI=0.6×min(54.65,150)+0.4×70=0.6×54.65+28=32.79+28=60.79

[0099] System determination: According to the preset grading criteria:

[0100]

[0101] The current SRI = 60.79 falls into the low-risk interval, triggering a first-level warning, and simultaneously recording the timestamp and physiological parameters into the database.

[0102] In summary, the present invention uses an infrared thermal imager to collect infrared thermal imaging images of both sides of the user's spine during sleep, uses temperature sensors to collect accurate temperature data of both sides of the user's spine during sleep, and uses pressure sensors to collect pressure data of the user acting on the pillow body during sleep. The control board calculates through the Synthetic Risk Index SRI, generates a stroke monitoring and warning analysis result, and sends the result to the mobile user terminal. The present application can realize the normalization and householdization of stroke detection, so that users can know their own stroke risk information in advance, and users can go to the hospital for professional examination and treatment based on the risk warning information.

[0103] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not intended to limit the patent application scope of the present invention. All equivalent changes or modifications completed under the technical spirit disclosed by the present invention shall fall within the patent scope covered by the present invention.

Claims

1. A stroke monitoring neck pillow system based on infrared thermal imaging, characterized in that, It includes a medical memory foam pillow body, an ultra-thin polyester fiber fabric pillowcase, a silicone cover, and a stroke detection integrated device. The stroke detection integrated device includes an infrared thermal imager, a temperature sensor, a pressure sensor, and a control board. The infrared thermal imager, temperature sensor, and pressure sensor are respectively connected to the control board. With the vertical centerline of the pillow body as a reference, the control board, pressure sensor, temperature sensor, and infrared thermal imager are arranged sequentially from top to bottom. The control board is located in the upper cavity inside the pillow body, the pressure sensor is located in the middle cavity inside the pillow body, the two temperature sensors are respectively located on the surface of the pillow body and are symmetrically arranged with respect to the vertical centerline of the pillow body, and the two infrared thermal imagers are respectively located on the surface of the pillow body and are symmetrically inclined with respect to the vertical centerline of the pillow body. The silicone sleeves cover the pressure sensor and the control board, and the pillowcase covers the infrared thermal imager, the temperature sensor, and the pillow body. An infrared thermal imager is used to acquire infrared thermal images of both sides of the spine during sleep. Temperature sensors are used to acquire precise temperature data of both sides of the spine during sleep, and pressure sensors are used to acquire pressure data on the pillow during sleep. The control board processes and analyzes the infrared thermal images, extracts the detected temperature data of both sides of the spine, and combines this with the precise temperature data to perform temperature analysis and determine the temperature difference between the two sides of the spine. ; The control panel is used to calculate the longest PDT (Prolonged Sleep Time) for a user in the same sleeping position based on pressure data, combined with the temperature difference between the two sides of the spine during sleep. Calculate the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI); The control board is used to determine the stroke risk level based on the user's Comprehensive Risk Index (SRI) during sleep, generate stroke monitoring and early warning analysis results, and send the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display. The postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI) are calculated as follows: The formula for calculating the postural ischemia index (PII) is as follows: ； The formula for calculating the Narrow Temperature Index (STI) is: ; The formula for calculating the SRI (Society of Risks Index) is as follows: .

2. The stroke monitoring neck pillow system based on infrared thermal imaging according to claim 1, characterized in that, The control board integrates a main control chip, a voltage conversion unit, a download circuit, a charging circuit, a Bluetooth circuit, and a timer circuit. An external power supply is connected to the voltage conversion unit or the charging circuit. The main control chip is connected to the voltage conversion unit. The main control chip and the voltage conversion unit are respectively connected to the download circuit, the Bluetooth circuit, and the timer circuit.

3. The stroke monitoring neck pillow system based on infrared thermal imaging according to claim 2, characterized in that, The voltage conversion unit includes a 3.7V to 5V circuit, a 5V to 3.3V circuit, and a 5V to -5V circuit. The 3.7V to 5V circuit and the 5V to -5V circuit provide power to the infrared thermal imager and pressure sensor, while the 5V to 3.3V circuit provides power to the main control chip, download circuit, Bluetooth circuit, and temperature sensor.

4. The stroke monitoring neck pillow system based on infrared thermal imaging according to claim 2, characterized in that, The external power supply input is 3.7V, and the main control chip is an STM32F103C8T6 chip. The download circuit is connected to the main control chip through the CLK and DIO interfaces, and uses an ST-LINK programmer with Keil5 software for programming. The charging circuit is connected to the external power supply 3.7V through a slide switch, and uses a Type-C USB interface. The Bluetooth circuit is connected to the main control chip through the USART interface, and uses the Bluetooth specification BLEV5.3+BR+EDR.

5. The stroke monitoring neck pillow system based on infrared thermal imaging according to claim 1, characterized in that, The pressure sensor includes four strain gauges arranged in an array. The strain gauge wire grid size is 3×3mm, the strain gauge base size is 6.8×4.2mm, the strain gauge sensitivity coefficient is 2.0±1%, the strain gauge output voltage is -3V to +4V, and the strain gauge output voltage accuracy is 0.001V.

6. A stroke monitoring and early warning method based on infrared thermal imaging, applied to the stroke monitoring neck pillow system based on infrared thermal imaging as described in any one of claims 1 to 5, characterized in that, Includes the following steps: S1: The infrared thermal imager collects infrared thermal images of both sides of the user's spine during sleep, the temperature sensor collects precise temperature data of both sides of the user's spine during sleep, and the pressure sensor collects pressure data of the user on the pillow during sleep. S2: The control board processes and analyzes the infrared thermal imaging image, extracts the detected temperature data on both sides of the spine, and combines it with precise temperature data to perform temperature analysis and judgment, obtaining the temperature difference between the two sides of the spine. ; S3: The control panel calculates the longest PDT (Progressive Temperature Difference) in the same sleeping position based on pressure data, and combines this with the temperature difference between the two sides of the spine during sleep. Calculate the postural ischemia index (PII), stenosis temperature difference index (STI), and comprehensive risk index (SRI); S4: The control board determines the stroke risk level based on the user's Comprehensive Risk Index (SRI) during sleep, generates stroke monitoring and early warning analysis results, and sends the stroke monitoring and early warning analysis results to the mobile user terminal for storage and display.

7. The stroke monitoring and early warning method based on infrared thermal imaging according to claim 6, characterized in that, Step S2 includes: The infrared thermal imaging image is preprocessed by using Gaussian filtering to remove background noise and sensing errors. The thermal data in the infrared thermal imaging image is converted into temperature values, and the temperature range of a normal human body is selected to distinguish human body temperature from ambient temperature. The infrared thermal imaging image is segmented to separate the human body area from the environment area, and Canny edge detection is used to extract the human body contour from the infrared thermal imaging image. The position of the spine in the infrared thermal image is used as a reference line for division. The left and right sides are divided according to the relative positions of the brain and spine. The center line is inferred by the temperature change of the spine region in the infrared thermal image, and the position of the center line is determined by an algorithm based on temperature gradient. The temperature change of each column of the infrared thermal image is calculated to find the place with the most obvious temperature change. The infrared thermal image is divided into left and right sides according to the center line.

8. The stroke monitoring and early warning method based on infrared thermal imaging according to claim 7, characterized in that, Step S2 also includes: After dividing the area into left and right sections, extract the temperature data from the left and right sections respectively, and calculate the temperature distribution of the corresponding areas; calculate the difference in average temperature between the two sections. Determine if there is a large temperature difference between the two sides; if the average temperature difference is significant... A temperature >0.2℃ indicates a significant temperature difference between the two sides, requiring temperature difference calculation; the difference between the lowest temperatures on both sides of the spine is calculated. ; Using precise temperature data provided by a temperature sensor, the system helps determine the temperature regions on both sides; it checks for any low-temperature regions within the temperature sensor's measurement area; if the temperature sensor's value differs significantly from the infrared thermal imaging result, and the low-temperature region in the infrared image is not present in the temperature sensor's area, the temperature sensor's value helps pinpoint the exact location of the low-temperature region; if the low-temperature region is within the temperature sensor's measurement area, the temperature sensor's value is used to correct the temperature value in the infrared thermal imaging image to improve the accuracy of the temperature difference calculation; finally, it outputs the temperature difference between the two sides of the spine. result.

9. The stroke monitoring and early warning method based on infrared thermal imaging according to claim 8, characterized in that, The control panel calculates the longest PDT (Prolonged Sleep Time) for the user in the same sleeping position based on pressure data, specifically: Set the pressure values ​​of the four strain gauges to be... Before analyzing the turning-over behavior, all stress data were normalized as follows: ， ； In the formula, strain gauge Pressure value; strain gauges after normalization Pressure value; This represents the maximum pressure value of the strain gauge. This is the minimum pressure value of the strain gauge; When monitoring whether a user turns over, the user's sleep behavior is divided into static state and dynamic state. The static state indicates that the user has not turned over, and the dynamic state indicates that the user has turned over. By calculating the rate of change of the pressure values ​​of the four strain gauges, it can be determined whether the user is currently in a dynamic state. The formula for calculating the rate of change of pressure is: ; In the formula, strain gauge The rate of change of the pressure value; After normalization Time-lapse strain gauge Pressure value; After normalization Time-lapse strain gauge Pressure value; Calculate the pressure changes of all strain gauges, and then calculate the sum of the pressure changes of all strain gauges. This sum will be used as an indicator of the user's recovery. The calculation formula is as follows: ; In the formula, The sum of the rates of change of pressure values ​​of all strain gauges; if If the value exceeds the preset threshold, the user is considered to have turned things around. The timing starts when pressure occurs, stops each time the user turns over, and records the duration. A new timing begins, with the subsequent duration overwriting the previous one, but not stored. The timing ends when the pressure reaches 0, yielding the longest physical time (PDT) for the user in the same sleeping position for each time period.

Images