A method for diabetes risk assessment based on multispectral night artificial light data

Abstract

This invention provides a method for diabetes risk assessment based on multispectral nighttime artificial light data, comprising: identifying the modeling population to be assessed and collecting data on the modeling population; performing spatial matching and geocoding on the modeling population to construct a spatial buffer for environmental exposure extraction and constructing structured indicators; acquiring multispectral nighttime artificial light remote sensing data covering the spatial buffer and preprocessing the multispectral nighttime artificial light remote sensing data; performing nighttime lighting type discrimination based on spectral ratio constraints on the preprocessed data, risk modeling and weight generation based on multispectral exposure response relationships, and employing heterogeneous risk assessment based on lighting type; fusing dual-channel nighttime artificial light risks to obtain a comprehensive nighttime artificial light diabetes risk index. This invention achieves quantitative assessment of diabetes risk through comprehensive processing of nighttime artificial light-related information.

Description

Technical Field

[0001] This invention relates to the field of remote sensing public health monitoring technology, and in particular to a method for assessing diabetes risk based on multispectral nighttime artificial light data. Background Technology

[0002] With rapid urbanization and infrastructure development, outdoor nighttime lighting has continued to grow globally, leading to widespread long-term exposure of the population to Artificial Light at Night (ALAN). Studies have shown that ALAN exposure is associated with various health problems, such as sleep disorders, mental health abnormalities, metabolic dysfunction, and an increased risk of certain chronic diseases. In China, the annual growth rate of ALAN has exceeded 6%, potentially posing a risk to residents' health. Currently, health risk assessment technologies based on nighttime light data still have the following limitations: (1) Lack of spectral information: Most existing studies on the relationship between ALAN and diabetes focus on light intensity indicators, usually characterizing light pollution exposure levels based on single-band nighttime light data, such as DMSP / OLS, VIIRS / DNB, or "Luojia-1". These data only reflect the overall radiation intensity of light and cannot distinguish the differences in the characteristics of different spectral components, thus failing to reveal the potential differential effects of different wavelengths of light on metabolic health risks. However, biomedical research shows that different wavelengths of light, especially short-wavelength blue light, have significantly different effects on the rhythm of melatonin secretion, circadian rhythm regulation, and glucose metabolism. Due to the lack of effective introduction of the spectral dimension of nighttime artificial light into current technologies, it is difficult to quantitatively assess the relative contribution of different spectral components to the formation of diabetes risk, thereby limiting the accuracy and specificity of health risk assessment results.

[0003] (2) Inability to identify lighting source types: With the rapid evolution of urban lighting technology, different types of lighting equipment coexist in the nighttime light environment, including white LEDs, yellow LEDs, red LEDs, high-pressure sodium lamps, and low-pressure sodium lamps. Different lighting technologies have significantly different spectral power distributions and radiation structures, and even under the same light intensity conditions, their potential impact on human physiological rhythms and metabolic processes also varies significantly. Existing health risk assessment methods based on nighttime light data usually treat nighttime artificial light as a homogeneous exposure factor, failing to effectively identify and distinguish different types of lighting sources. They also lack the technical means to combine lighting type information with health risk models, resulting in an inability to characterize the heterogeneous characteristics of nighttime artificial light exposure at the source structure level, thus restricting the level of precision in risk assessment.

[0004] (3) Insufficient multi-source data fusion: Existing methods for assessing the environmental risks of diabetes still fall short in terms of multi-source data fusion, making it difficult to simultaneously integrate multi-dimensional characteristic information such as the intensity, spectral composition, and type of artificial light at night within a unified framework. Due to the lack of multi-dimensional collaborative modeling methods for exposure to artificial light at night, existing assessment results often exhibit low spatial resolution, unclear risk sources, and difficulty in supporting refined health risk management and intervention decisions.

[0005] Therefore, there is an urgent need to propose a diabetes risk assessment technology that can integrate the multispectral characteristics of artificial light at night, lighting type information, and individual health characteristics, in order to overcome the shortcomings of existing technologies in the utilization of spectral information, differentiation of lighting types, and fusion of multi-source data, and to achieve a more refined assessment of the risk of metabolic diseases such as diabetes. Summary of the Invention

[0006] This invention provides a method for assessing the risk of diabetes based on multispectral nighttime artificial light data, which addresses the shortcomings of existing technologies. By comprehensively processing information related to nighttime artificial light, a quantitative assessment of the risk of developing diabetes can be achieved.

[0007] This invention provides a method for assessing diabetes risk based on multispectral nighttime artificial light data, comprising: Identify the target population for modeling and collect data on that population. Spatial matching and geocoding are performed on the modeled population to construct a spatial buffer for environmental exposure extraction, and structured indicators are constructed. Acquire multispectral nighttime artificial light remote sensing data covering the spatial buffer zone, and preprocess the multispectral nighttime artificial light remote sensing data; The preprocessed data were used to identify the type of nighttime lighting based on spectral ratio constraints, and risk modeling and weight generation were performed based on multispectral exposure response relationships. Heterogeneous risk assessment based on lighting type was also adopted. By integrating the risks of artificial light at night from both channels, a comprehensive risk index for diabetes caused by artificial light at night was obtained.

[0008] According to the present invention, a method for assessing diabetes risk based on multispectral nighttime artificial light data is provided, which identifies the modeling population to be assessed, including: Elderly diabetic patients who meet the inclusion and exclusion criteria are collected according to inclusion criteria, which include having available spatial location information of residence and maintaining a stable residence within a preset time window to ensure spatial consistency of nighttime artificial light exposure.

[0009] According to the present invention, a method for diabetes risk assessment based on multispectral nighttime artificial light data is provided, which involves collecting data from a modeling population, including: Collect individual characteristic information of the modeled population, including demographic characteristics, health information, and environmental and behavioral exposure characteristics.

[0010] According to the present invention, a method for diabetes risk assessment based on multispectral nighttime artificial light data is provided, which involves spatial matching and geocoding of the modeled population to construct a spatial buffer for environmental exposure extraction, including: The residential addresses of the modeled population are geocoded to obtain their corresponding spatial coordinates; A spatial buffer for environmental exposure extraction is constructed with the aforementioned spatial location coordinates as the center.

[0011] According to the present invention, a method for assessing diabetes risk based on multispectral nighttime artificial light data is provided, which includes constructing structured indicators, including: Based on individual characteristic information, a structured feature index set for diabetes risk assessment is constructed, and the structured feature index set is standardized and spatially weighted.

[0012] According to the present invention, a method for diabetes risk assessment based on multispectral nighttime artificial light data is provided, comprising acquiring multispectral nighttime artificial light remote sensing data covering the spatial buffer, and preprocessing the multispectral nighttime artificial light remote sensing data, including: Radiometric calibration processing was performed on the multispectral nighttime artificial light remote sensing data to obtain the radiance values ​​of red, green, blue and panchromatic bands; Based on the radiance value, the multispectral nighttime artificial light remote sensing data is subjected to quality control processing targeting the nighttime low-light imaging characteristics, including abnormal pixel removal and point spread function correction. Based on the residential coordinates of the modeled population, multispectral radiance features within the corresponding spatial neighborhood are extracted from the multispectral nighttime artificial light remote sensing data to construct a nighttime artificial light multispectral feature dataset.

[0013] According to the present invention, a method for assessing diabetes risk based on multispectral nighttime artificial light data is provided, which involves determining the type of nighttime lighting based on spectral ratio constraints on preprocessed data, including: Based on the spectral radiance information of the red, green, and blue light bands, a spectral ratio feature is constructed. By combining spatial consistency constraints to identify nighttime lighting types, nighttime lighting type characteristics for diabetes risk assessment are obtained.

[0014] The present invention provides a method for assessing diabetes risk based on multispectral nighttime artificial light data, comprising risk modeling and weight generation based on multispectral exposure-response relationships, including: Based on the individual characteristics of the modeling population and the multispectral exposure information of artificial light at night, a multi-level exposure-response relationship model is constructed. A spectral risk weight is generated for nighttime artificial light exposure assessment through a hierarchical correction and risk response mapping mechanism.

[0015] A diabetes risk assessment method based on multispectral nighttime artificial light data provided by the present invention employs a heterogeneous risk assessment based on lighting type, including: The obtained nighttime lighting type characteristics are introduced as conditional variables into the diabetes risk assessment model. Individual demographic, behavioral, and health characteristics are fixed, and the differences in health risks introduced by different lighting technology paths are quantified to achieve independent identification and assessment of the heterogeneity effect of nighttime artificial light lighting types.

[0016] According to the present invention, a method for assessing diabetes risk based on multispectral nighttime artificial light data is provided, which integrates dual-channel nighttime artificial light risk to obtain a comprehensive risk index for diabetes caused by nighttime artificial light, including: The constructed risk index for exposure to artificial light at night was weighted and fused with the constructed risk index for exposure to lighting type to obtain the final comprehensive risk index for diabetes caused by artificial light at night.

[0017] This invention provides a diabetes risk assessment method based on multispectral nighttime artificial light data. By introducing multispectral nighttime artificial light data and modeling and weighting different spectral components such as red, green, and blue light, it can identify the differentiated impact of different spectral nighttime artificial light on diabetes risk, overcoming the deficiency of existing technologies that ignore nighttime spectral characteristics. This invention distinguishes the health risks of nighttime artificial light under different lighting conditions by differentiating nighttime lighting types and conducting differentiated risk assessments based on these types, helping to reveal the potential impact mechanisms of different lighting types on diabetes risk. This invention comprehensively considers demographic characteristics, health-related characteristics, and environmental and behavioral exposure factors during the risk assessment process, performing multi-level parameter correction on the model to effectively reduce the impact of confounding factors on the assessment results and improve the robustness of the risk assessment results. This invention uses nighttime light remote sensing data for risk assessment, without relying on individual-worn devices or long-term field monitoring. It has the advantages of wide coverage, low acquisition cost, and high repeatability. The assessment results can be further used for risk level display and spatial distribution, making it suitable for large-scale population diabetes risk assessment and nighttime light environment health management applications. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is one of the flowcharts of the diabetes risk assessment method based on multispectral nighttime artificial light data provided by the present invention; Figure 2 This is the second flowchart of the diabetes risk assessment method based on multispectral nighttime artificial light data provided by the present invention; Figure 3 This is a flowchart of the data collection and indicator construction process for the modeling population provided by the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] Figure 1 This is one of the flowcharts illustrating the diabetes risk assessment method based on multispectral nighttime artificial light data provided in this embodiment of the invention, such as... Figure 1 As shown, it includes: Step 100: Identify the target population for modeling and collect data on the target population; Step 200: Perform spatial matching and geocoding on the modeled population, construct a spatial buffer for environmental exposure extraction, and construct structured indicators; Step 300: Acquire multispectral nighttime artificial light remote sensing data covering the spatial buffer zone, and preprocess the multispectral nighttime artificial light remote sensing data; Step 400: Perform nighttime lighting type discrimination based on spectral ratio constraints on the preprocessed data, risk modeling and weight generation based on multispectral exposure response relationship, and adopt heterogeneity risk assessment based on lighting type; Step 500: Integrate the dual-channel nighttime artificial light risk to obtain a comprehensive risk index for diabetes caused by nighttime artificial light.

[0022] Specifically, embodiments of the present invention include, for example... Figure 2 As shown, it includes the following steps: S1: Selection of the modeling population: Elderly diabetic patients meeting the inclusion and exclusion criteria were collected. The inclusion criteria were: (1) Age ≥ 45 years; (2) Long-term residents of the survey area (who have lived in the area for more than 6 months prior to the survey); (3) Has not participated in clinical trials or other intervention programs in the past 3 months.

[0023] S2: Data Collection for the Modeling Population: Individual characteristic information of the modeling population is collected through questionnaires and on-site measurements, including demographic characteristics, health information, and environmental and behavioral exposure characteristics. Specifically, this includes: (1) Demographic characteristics: age, sex, marital status, education level, and family income; (2) Health information: height, weight, blood pressure; (3) Environmental and behavioral exposure characteristics: physical activity level, smoking status, alcohol consumption status, exposure level of green space around residence, and air pollution exposure level.

[0024] S3: Spatial Matching and Geocoding: Obtain the residential addresses of the modeling population and use the Baidu Maps Coding API service to geocode the participants' addresses to the corresponding longitude and latitude.

[0025] S4: Structured Indicator Construction: Based on the individual characteristic information, a set of structured feature indicators for diabetes risk assessment is constructed. These indicators are then standardized and spatially weighted to improve the comparability and stability between different features. Figure 3 As shown, it specifically includes: (1) Classify and encode discrete variables such as gender, marital status, smoking status and drinking status.

[0026] (2) Classify and code the education level and family income; among them: education level is divided into 0-6 years, 7-12 years and >12 years; family income is divided into three categories: <20,000 yuan per year, 20,000-39,999 yuan and >39,999 yuan per year.

[0027] (3) Calculate the Body Mass Index (BMI) based on weight and height information. The calculation formula is as follows: BMI = weight (kg) / height² (m²) Based on preset threshold ranges, BMI is divided into underweight (below 18.5), normal (18.5-23.9), overweight (24-27.9), and obese (≥28).

[0028] (4) Based on blood pressure measurement values ​​and previous hypertension history information, construct blood pressure status judgment rules; when systolic or diastolic blood pressure exceeds the preset threshold, or there is a clear history of hypertension, it is marked as a hypertension state, otherwise it is marked as a non-hypertension state, thus forming a binary health risk indicator.

[0029] (5) The physical activity scale of the elderly is used to assess the individual's physical activity level. Combined with information on activity frequency and activity intensity, it is determined whether the individual meets the recommended standards of the guidelines. The physical activity status is coded as low activity exposure or standard activity exposure to characterize the metabolic regulation capacity at the behavioral level.

[0030] (6) Construction of spatially weighted green space exposure index: Based on the normalized vegetation index (NDVI) remote sensing image data, a 30-meter spatial buffer zone is constructed with the individual residence as the center; within the buffer zone, the NDVI pixel values ​​are spatially aggregated and calculated to obtain the weighted NDVI exposure index, which is used to characterize the comprehensive exposure level of the green space environment around the individual.

[0031] (7) Construction of air pollution exposure index based on spatial statistics: Based on PM2.5 remote sensing image data, a 1km spatial buffer zone is constructed with the individual residence as the center; the annual average concentration of PM2.5 is calculated within the buffer zone, and the interference of outliers is reduced by spatial smoothing to obtain a stable air pollution exposure index.

[0032] (8) Feature standardization and indicator integration: The various structured feature indicators constructed above are uniformly standardized to form an input feature matrix for subsequent diabetes risk modeling.

[0033] S5: Nighttime Artificial Light Data Preprocessing: Acquire multispectral nighttime artificial light remote sensing data covering the aforementioned spatial location, and perform radiometric calibration and quality control processing on the nighttime artificial light remote sensing data to obtain spectral radiance information for red, green, blue, and panchromatic bands for subsequent risk assessment. Specifically, this includes: (1) Download the SDGSAT-1 Low Light Imager (GLI) L4A level nighttime artificial light data covering the residential addresses of the modeled population. The SDGSAT-1 satellite's GLI is specifically designed to detect weak nighttime light radiation, providing high-precision urban light data. It features panchromatic (PH) and red (R), green (G), and blue (B) color observation modes, with a spatial resolution of 10 meters for the panchromatic band and 40 meters for the color band. GLI data products include three standard levels: L1, L2, and L4. The L1 level product is a standard product generated based on the L0 level data after relative radiometric correction, band registration, and HDR fusion. The L2 level product is generated based on the L1 level data after geometric correction. The L4 level product is generated by orthorectifying the L1 level data using ground control points and a digital elevation model, and output in a standardized format. Currently, only the L4 level product is available to users.

[0034] (2) Radiometric calibration is performed on the GLI data to convert the remote sensing image count values ​​into spectral radiance with clear physical meaning. The calculation formula is as follows:

[0035] Where L is the spectral radiance, with units of W / (m²). 2 / sr / μm); DN represents the count value of the image after relative radiometric calibration; This is the radiation calibration gain coefficient. This is the radiometric calibration bias term, and its value is derived from the SDGSAT-1 satellite user manual, as shown in Table 1.

[0036] Table 1 Absolute radiometric calibration coefficients of the SDGSAT-1 satellite low-light payload

[0037] (3) Based on radiometric calibration, quality control processing targeting the low-light imaging characteristics of nighttime artificial light is performed on the multispectral nighttime artificial light remote sensing data, including abnormal pixel removal and point spread function correction, to improve the authenticity and stability of nighttime spectral radiation information in health risk assessment scenarios. Specifically, based on preset brightness thresholds and time stability criteria, low-brightness noise pixels, transient light source pixels, and unstable luminous areas affected by system noise are identified and removed; further, a point spread function correction operator is introduced, and the scattering kernel function caused by atmospheric scattering and system response during nighttime imaging is estimated using the dark pixel distribution characteristics of non-luminous background areas in the study area; based on the scattering kernel function, deconvolution processing is performed on the multispectral nighttime artificial light image to weaken the brightness overestimation caused by the diffusion of high-brightness light sources to the surrounding areas, thereby restoring the remote sensing observation brightness to a spectral radiation brightness value that is closer to the real ground radiation characteristics; through the above processing, the relative deviation caused by scattering effects between different spectral bands can be effectively reduced, the reliability of spectral ratio calculation and lighting type discrimination can be improved, and high-reliability input data can be provided for subsequent nighttime artificial light health risk modeling.

[0038] (4) Based on the residential coordinates of the modeled population, extract the multispectral radiance features within the corresponding spatial neighborhood of the nighttime artificial light image to form a nighttime artificial light multispectral feature dataset oriented towards individual exposure scale.

[0039] S6: Nighttime Lighting Type Discrimination Based on Spectral Ratio Constraints: Based on the spectral radiance information of the red, green, and blue light bands, spectral ratio features are constructed, and combined with spatial consistency constraints to discriminate nighttime lighting types, resulting in lighting type features for diabetes risk assessment. Specifically, this includes: (1) Obtain the spectral radiance values ​​of the red (R), green (G), and blue (B) bands of the corresponding regions, calculate the intensity ratio of blue to green (B / G) and the intensity ratio of green to red (G / R), and construct the spectral ratio characteristics of artificial light at night. The intensity ratio ranges corresponding to different lighting types are shown in Table 2. "B / G" is the ratio of blue to green light, and "G / R" is the ratio of green to red light. "-" indicates no boundary limit. Pixels that do not conform to the standards of the other four lighting categories are also classified as YLED.

[0040] (2) Based on the preset spectral ratio threshold rule, the lighting type of a single pixel is initially determined, and lighting types such as white LED, yellow LED, red LED, high pressure sodium lamp (HPS) and low pressure sodium lamp (LPS) are identified; (3) Based on the pixel-level discrimination results, a spatial consistency constraint mechanism is introduced. Within the preset spatial neighborhood around the residence, the lighting type discrimination results are statistically summarized to determine the dominant lighting type in the area, thereby reducing the impact of single-pixel misjudgment on individual exposure assessment results. (4) The dominant lighting type is used as a structured nighttime artificial light exposure feature and input into the subsequent diabetes risk assessment model to characterize the heterogeneity of nighttime artificial light exposure under different lighting technology conditions, as shown in Table 2.

[0041] Table 2 Classification Standards for Lighting Types

[0042] S7: Risk Modeling and Weight Generation Based on Multispectral Exposure-Response Relationships: Based on individual characteristics of the modeling population and multispectral exposure information to artificial light at night, a multi-level exposure-response relationship model is constructed. Through hierarchical correction and risk-response mapping mechanisms, spectral risk weights are generated for assessing exposure to artificial light at night. Specifically, this includes: (1) Using whether an individual has diabetes as the dependent variable, and the brightness of red light (R), green light (G), blue light (B) and panchromatic light (P) at night as the main independent variables, and combined with individual characteristic information, a logistic regression model was constructed to characterize the relationship between nighttime artificial light exposure and diabetes risk. (2) The logistic regression model is subjected to multi-level progressive correction to gradually remove the interference of demographic factors, environmental behavioral factors and individual health characteristics on the nighttime artificial light exposure effect, including: Model 1 (Basic Correction): Only age and gender are corrected.

[0043] Model 2 (Environmental Behavior Correction): Based on Model 1, add marital status, education level, family income, physical activity, smoking status, drinking status, green space near residence, and air pollution.

[0044] Model 3 (Full Health Characteristics Correction): Based on Model 2, BMI and blood pressure are added.

[0045] Through comparative analysis of multi-level models, spectral exposure parameters with consistent risk direction and stable effects under different correction levels were selected. (3) Based on the model parameters stabilized after multi-level correction, the odds ratio (OR) of each spectral band with the risk of diabetes was calculated, and the risk response intensity parameter of spectral exposure was constructed using its natural logarithm:

[0046] Where i∈{R, G, B, P}; Characterizes the relative response intensity of nighttime artificial light in the i-th spectral band to the risk of diabetes; (4) Input the risk response intensity parameter into a preset weight mapping function, and convert the risk response intensity into a basic weight Wi for index construction based on the nonlinear characteristics of different spectral exposure effects. The mapping function includes, but is not limited to, linear normalization, logarithmic mapping, exponential mapping or hierarchical mapping based on domain experience.

[0047] (5) Based on the generated spectral weights, the nighttime artificial light radiance of each spectrum is weighted and integrated to construct a nighttime artificial light spectral exposure risk index. This is used to characterize the combined effects of different spectral compositions on diabetes risk.

[0048]

[0049] in, As an indicator of nighttime exposure risk to artificial spectra, This refers to the spectral radiance of the band corresponding to step S5.

[0050] S8: Heterogeneity risk assessment based on lighting type: The nighttime lighting type characteristics obtained in step S6 are introduced as conditional variables into the diabetes risk assessment model. Under the premise of fixed individual demographic characteristics, behavioral characteristics, and health characteristics, the differences in health risks introduced by different lighting technology paths are quantified, thereby achieving independent identification and assessment of the heterogeneity effect of nighttime artificial light lighting type.

[0051] (1) Using a logistic regression model, while keeping other covariates constant, calculate the ratio corresponding to the lighting type k described in S6. ), and extract its risk impact coefficient. :

[0052] Among them, k∈{WLED, YLED, RLED, HPS, LPS}; is the original weighting coefficient for the k-th lighting type.

[0053] (2) To eliminate the differences in the dimensions of parameters from different models and enhance the comparability of indicators, the original risk impact coefficients were adjusted. By performing a normalized mapping function, the risk weights for different lighting types can be obtained. The mapping methods include, but are not limited to, linear normalization, piecewise normalization, or proportional mapping based on empirical constraints:

[0054] in, Characterize the relative contribution of different lighting types to the risk of diabetes under the same light intensity conditions.

[0055] (3) Based on the dominant lighting type and its corresponding total nighttime artificial light radiance identified within the participants' residential area. Calculate the risk index of exposure to artificial light types at night. : ) S9: Dual-channel risk fusion of nighttime artificial light: To simultaneously characterize the health risk contribution of nighttime artificial light in both the spectral composition and illumination type dimensions, a dual-channel risk fusion mechanism is introduced to fuse the nighttime artificial light spectral exposure risk index constructed in step S7. The lighting type exposure risk index constructed in step S8 Weighted fusion is performed to obtain the final comprehensive index of diabetes risk from artificial light at night. .

[0056]

[0057] in, and These are the contribution coefficients for spectral characteristics and illumination type, respectively. In specific implementations, these contribution coefficients can be set based on sample goodness of fit, model stability evaluation results, or expert experience. In this embodiment, let... and Both are set to 0.5 to achieve a balanced and comprehensive evaluation of spectral effects and illumination type effects.

[0058] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0059] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for diabetes risk assessment based on multispectral night-time artificial light data, characterized in that, The method comprises the following steps: determining a modeling population to be evaluated, collecting data of the modeling population; spatial matching and geocoding the modeling population, constructing a spatial buffer for environmental exposure extraction, and constructing structured indicators; acquiring multispectral night artificial light remote sensing data covering the spatial buffer, and preprocessing the multispectral night artificial light remote sensing data; performing night lighting type discrimination based on spectral ratio constraints on the preprocessed data, risk modeling and weight generation based on multispectral exposure response relationships, and adopting heterogeneous risk assessment based on lighting types; fusing double-channel night artificial light risks to obtain a night artificial light diabetes comprehensive risk indicator.

2. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 1, characterized in that, determining a modeling population to be evaluated, comprising: collecting elderly diabetic patients meeting the inclusion and exclusion criteria according to the inclusion conditions, wherein the inclusion conditions include having available spatial location information of the residence and the residence remaining stable within a preset time window, so as to ensure the spatial consistency of night artificial light exposure.

3. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 2, characterized in that, collecting data of the modeling population, comprising: collecting individual characteristic information of the modeling population, including demographic characteristics, health information, environmental and behavioral exposure characteristic information.

4. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 1, characterized in that, spatial matching and geocoding the modeling population, constructing a spatial buffer for environmental exposure extraction, comprising: geocoding the residence address of the modeling population to obtain corresponding spatial position coordinates; constructing a spatial buffer for environmental exposure extraction with the spatial position coordinates as the center.

5. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 4, characterized in that, constructing structured indicators, comprising: based on individual characteristic information, constructing a structured characteristic indicator set for diabetes risk assessment, and performing standardization processing and spatial weighting calculation on the structured characteristic indicator set.

6. The method of diabetes risk assessment based on multispectral night-time artificial light data according to claim 1, characterized in that, acquiring multispectral night artificial light remote sensing data covering the spatial buffer, and preprocessing the multispectral night artificial light remote sensing data, comprising: performing radiometric calibration processing on the multispectral night artificial light remote sensing data to obtain radiometric brightness values of red, green, blue and panchromatic bands; based on the radiometric brightness values, performing quality control processing on the multispectral night artificial light remote sensing data for night micro-light imaging characteristics, including abnormal pixel removal and point spread function correction; based on the residence coordinates of the modeling population, extracting multispectral radiometric brightness features within the corresponding spatial neighborhood range in the multispectral night artificial light remote sensing data to construct a night artificial light multispectral feature data set.

7. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 6, characterized in that, performing night lighting type discrimination based on spectral ratio constraints on the preprocessed data, comprising: based on the spectral radiometric brightness information of the red, green and blue bands, constructing spectral ratio features; combining spatial consistency constraints to discriminate night lighting types to obtain night lighting type features for diabetes risk assessment.

8. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 7, characterized in that, risk modeling and weight generation based on multispectral exposure response relationships, comprising: based on individual characteristic information of the modeling population and night artificial light multispectral exposure information, constructing a multi-level exposure-response relationship model; through hierarchical correction and risk response mapping mechanism, generating spectral risk weights for night artificial light exposure assessment.

9. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 8, characterized in that, Heterogeneity risk assessment based on lighting type is employed, including: The obtained nighttime lighting type characteristics are introduced as conditional variables into the diabetes risk assessment model. Individual demographic, behavioral, and health characteristics are fixed, and the differences in health risks introduced by different lighting technology paths are quantified to achieve independent identification and assessment of the heterogeneity effect of nighttime artificial light lighting types.

10. The method for diabetes risk assessment based on multispectral night artificial light data according to claim 1, characterized in that, By integrating the risks of dual-channel artificial light at night, a comprehensive risk index for diabetes caused by artificial light at night was obtained, including: The constructed risk index for exposure to artificial light at night was weighted and fused with the constructed risk index for exposure to lighting type to obtain the final comprehensive risk index for diabetes caused by artificial light at night.

Images