CN-121977704-A - Indoor thermal environment dynamic assessment method and system based on artificial intelligence

CN121977704ACN 121977704 ACN121977704 ACN 121977704ACN-121977704-A

Abstract

The invention relates to the technical field of thermal environment assessment and discloses an indoor thermal environment dynamic assessment method and system based on artificial intelligence, wherein the method comprises the steps of performing temperature compensation on an infrared thermal imaging sequence in preprocessed multi-source thermal sensing data by using a temperature compensation algorithm based on time-frequency domain joint analysis to obtain a pure infrared thermal imaging sequence, and extracting an environmental temperature time sequence characteristic of the environmental thermal sensing data sequence in the preprocessed multi-source thermal sensing data; performing multi-scale skin temperature decomposition and enhancement treatment on the pure infrared thermal imaging sequence, and extracting time sequence change characteristics of skin temperature; and dynamically evaluating the indoor thermal environment by using a thermal environment evaluation model fused with skin temperature perception. According to the invention, through temperature compensation, multi-scale skin temperature enhancement and skin temperature characteristic extraction on infrared thermal imaging, a thermal environment assessment model which integrates skin temperature sensing and environmental temperature time sequence characteristics is adopted to realize dynamic assessment of indoor thermal environment.

Inventors

WANG SHILIANG
WANG CHENGLIN
LIU QIANG
XU YUHAO
LI YANWEI
YU JUAN
LIU XIANG
Tan Lingye

Assignees

济南大学

Dates

Publication Date: 20260505
Application Date: 20260120

Claims (9)

1. An artificial intelligence-based indoor thermal environment dynamic assessment method is characterized by comprising the following steps: s1, collecting multi-source heat sensing data of an indoor environment, and performing time synchronization and missing value processing on the multi-source heat sensing data to obtain preprocessed multi-source heat sensing data; s2, according to the preprocessed multi-source heat sensing data, performing temperature compensation on an infrared thermal imaging sequence in the preprocessed multi-source heat sensing data by using a temperature compensation algorithm based on time-frequency domain joint analysis to obtain a pure infrared thermal imaging sequence, and extracting the environmental temperature time sequence characteristics of an environmental heat sensing data sequence in the preprocessed multi-source heat sensing data; S3, performing multi-scale skin temperature decomposition and enhancement treatment on the pure infrared thermal imaging in the pure infrared thermal imaging sequence to obtain an enhanced pure infrared thermal imaging sequence, and extracting time sequence change characteristics of skin temperature in the enhanced pure infrared thermal imaging sequence; S4, dynamically evaluating the indoor thermal environment by using a thermal environment evaluation model fused with skin temperature perception according to the time sequence characteristics of the environmental temperature and the time sequence change characteristics of the skin temperature.
2. The method for dynamically evaluating indoor thermal environment based on artificial intelligence according to claim 1, wherein the step S1 of collecting multi-source thermal sensing data of indoor environment, performing time synchronization and missing value processing on the multi-source thermal sensing data, comprises: S11, deploying a plurality of types of heat sensors in an indoor environment, and periodically acquiring multi-source heat sensing data of the indoor environment by using the heat sensors, wherein the multi-source heat sensing data comprise an infrared heat imaging sequence of the surface of a human body and an environment heat sensing data sequence, and the environment heat sensing data sequence comprises an environment wind speed sequence and an environment humidity sequence; S12, constructing a unified time axis of the multi-source heat sensing data, and extracting the acquired data of time nodes of the multi-source heat sensing data in the unified time axis to obtain multi-source heat sensing data with synchronous time; And S13, traversing the time-synchronous multi-source heat sensing data, taking the acquired data missing from the time nodes as missing values, and complementing the missing values by adopting a linear interpolation mode to obtain the preprocessed multi-source heat sensing data.
3. The method for dynamically evaluating indoor thermal environment based on artificial intelligence according to claim 1, wherein in step S2, the temperature compensation algorithm based on the time-frequency domain joint analysis is used to perform temperature compensation on the infrared thermal imaging sequence in the preprocessed multi-source thermal sensing data, and the method comprises the following steps: s21, performing time domain discrete Fourier transform on each pixel point in infrared thermal imaging in the infrared thermal imaging sequence according to the infrared thermal imaging sequence in the preprocessed multi-source thermal sensing data to obtain a time spectrum of the pixel point; S22, carrying out frequency domain smoothing filtering on the time spectrum of the pixel point by utilizing a frequency domain filter to obtain a smoothed time spectrum of the pixel point; S23, carrying out image reconstruction on the infrared thermal imaging according to the smooth time spectrum of the pixel points to obtain an infrared thermal imaging sequence and an infrared thermal imaging sequence after time-frequency domain smoothing; s24, respectively extracting low-frequency effective components of an ambient wind speed sequence and an ambient humidity sequence according to an ambient heat sensing data sequence in the preprocessed multi-source heat sensing data; S25, generating an ambient temperature compensation term of the infrared thermal imaging after the time-frequency domain smoothing according to the low-frequency effective components of the ambient wind speed sequence and the ambient humidity sequence, performing temperature compensation on the infrared thermal imaging after the time-frequency domain smoothing to obtain pure infrared thermal imaging, and sequencing and splicing the pure infrared thermal imaging according to the sequence of time nodes associated with the pure infrared thermal imaging to obtain a pure infrared thermal imaging sequence.
4. The method for dynamically evaluating indoor thermal environment based on artificial intelligence according to claim 3, wherein the step S2 of extracting the environmental temperature time sequence characteristics of the environmental thermal sensing data sequence in the preprocessed multi-source thermal sensing data further comprises: S26, generating a wind speed-temperature contribution coefficient and a humidity-temperature contribution coefficient according to the pure infrared thermal imaging sequence; S27, according to the low-frequency effective components of the environment wind speed sequence and the environment humidity sequence, converting the low-frequency effective components into wind speed contribution items and humidity contribution items of different time nodes by utilizing a wind speed-temperature contribution coefficient and a humidity-temperature contribution coefficient, and taking the sum of the wind speed contribution items and the humidity contribution items as an environment temperature contribution item at the time node; And S28, sequencing the environmental temperature contribution items at the time nodes according to the sequence of the time nodes to obtain an environmental temperature contribution item sequence, and performing first-order differential operation on the environmental temperature contribution item sequence to obtain the environmental temperature time sequence characteristics of the environmental heat sensing data sequence.
5. The method for dynamically estimating indoor thermal environment based on artificial intelligence according to claim 3, wherein the formula for performing frequency domain smoothing filtering on the time spectrum of the pixel point by using a frequency domain filter in the step S22 is as follows: ; ; Wherein, the Representing a time spectrum The corresponding smoothed time spectrum is then used to produce, Representing pixel points in the infrared thermal imaging In the first place The time spectrum of the time-frequency index, , Representing the number of pixel rows for infrared thermal imaging, Representing the number of columns of pixels, and the pixels of an infrared thermal image Representing the pixel points of the x-th row and the y-th column in the infrared thermal imaging, N represents the length of the infrared thermal imaging sequence, Representing the frequency domain filter, Indicating that the frequency filter is for the first The filter coefficients of the time-frequency index, Representing a time spectrum Is used for the time-frequency of the actual time frequency of (c), Representing the cut-off frequency in the frequency domain filter, Representing the coefficient of smoothing and the coefficient of smoothing, Representing the time interval between adjacent time nodes in the unified time axis.
6. The method for dynamically evaluating indoor thermal environment based on artificial intelligence according to claim 1, wherein the step S3 of performing multi-scale skin temperature decomposition and enhancement processing on the pure infrared thermal images in the pure infrared thermal imaging sequence comprises: S31, carrying out Laplacian pyramid decomposition on the pure infrared thermal imaging to obtain a low-frequency basal layer and a plurality of high-frequency detail layers of the pure infrared thermal imaging, wherein the low-frequency basal layer reflects the overall thermal distribution trend of the skin, and the high-frequency detail layers reflect the gradient change of local temperature; S32, generating skin density weight of the pixel points at the positions corresponding to the low-frequency basal layer and the high-frequency detail layer based on the pixel values of the pixel points in the pure infrared thermal imaging; S33, utilizing the skin density weight to carry out weighted modulation on the low-frequency basal layer and the high-frequency detail layer to obtain the low-frequency basal layer and the high-frequency detail layer after weighted modulation; And S34, carrying out fusion reconstruction on the low-frequency basal layer and the high-frequency detail layer after weighted modulation according to the Laplacian pyramid reconstruction sequence, generating enhanced pure infrared thermal imaging corresponding to the pure infrared thermal imaging, and sequencing the enhanced pure infrared thermal imaging according to the time node sequence to obtain an enhanced pure infrared thermal imaging sequence.
7. The method for dynamically evaluating indoor thermal environment based on artificial intelligence according to claim 6, wherein the step S3 of extracting the time-series variation characteristic of skin temperature in the enhanced pure infrared thermal imaging sequence further comprises: S35, calculating a pixel value mean value and a pixel value standard deviation of the enhanced pure infrared thermal imaging according to the enhanced pure infrared thermal imaging sequence to form a pixel value mean value sequence and a pixel value standard deviation sequence; S36, performing first-order differential operation on the pixel value mean value sequence and the pixel value standard deviation sequence respectively to obtain a first-order differential sequence of the pixel value mean value and a first-order differential sequence of the pixel value standard deviation, and taking the first-order differential sequence and the first-order differential sequence of the pixel value standard deviation as time sequence change characteristics of skin temperature in the enhanced pure infrared thermal imaging sequence.
8. The method for dynamically estimating indoor thermal environment based on artificial intelligence according to claim 1, wherein the step S4 of dynamically estimating indoor thermal environment by using a fused skin temperature-aware thermal environment estimation model comprises the steps of: the thermal environment assessment model comprises a characteristic input layer, a time sequence characteristic coding layer, a thermal comfort level regression layer and a result convergence layer, and the dynamic assessment flow of the indoor thermal environment is as follows: S41, receiving the time sequence change characteristics of the skin temperature and the time sequence characteristics of the ambient temperature in the enhanced pure infrared thermal imaging sequence by the characteristic input layer according to the time sequence change characteristics of the ambient temperature and the time sequence change characteristics of the skin temperature; S42, the time sequence feature coding layer performs time sequence coding on the time sequence change feature and the environment temperature time sequence feature to obtain a hidden state coding vector corresponding to any time node; s43, mapping the hidden state coding vector into a thermal comfort level evaluation value by a thermal comfort level regression layer, and taking the thermal comfort level evaluation value as an indoor thermal environment evaluation result corresponding to a time node; and S44, the result convergence layer performs time sequence weighting on the indoor thermal environment evaluation results of all the time nodes to obtain a comprehensive evaluation result of the indoor thermal environment.
9. An artificial intelligence-based indoor thermal environment dynamic assessment system, which comprises a data processing module, a temperature compensation module, a feature extraction module and a dynamic assessment module, so as to realize an artificial intelligence-based indoor thermal environment dynamic assessment method according to any one of claims 1-8.

Description

Indoor thermal environment dynamic assessment method and system based on artificial intelligence Technical Field The invention relates to the technical field of big data processing, in particular to the field of thermal environment assessment, and specifically relates to an indoor thermal environment dynamic assessment method and system based on artificial intelligence. Background With the rapid development of intelligent building technology and the wide application of the internet of things technology, real-time dynamic monitoring and intelligent regulation of indoor thermal environment have become key problems in modern building research and application. Accurate thermal environment management not only can promote the comfort level of resident, improve life and work experience, but also plays an important role in building energy conservation, energy consumption reduction and environmental sustainability. In addition, the requirements of health management on indoor thermal environments are increasingly prominent, and especially in living environments of old people, children and high-risk people, reasonable control of temperature and humidity is of great significance in maintaining physiological steady state and reducing heat stress response. The traditional air conditioner and environment regulation method mainly depends on a fixed temperature or humidity threshold value for regulation, and the mode cannot consider individual difference and subjective heat sensation change of a human body, so that imbalance of indoor local heat environment, large temperature and humidity fluctuation and increased energy consumption are usually caused even when the air conditioner operates under high load. In addition, the fixed threshold regulation is difficult to respond to dynamically changing indoor thermal environments, such as personnel activities, indoor heat source changes or external climate fluctuations, so that thermal comfort assessment is lagged and inaccurate, and the overall experience and energy saving effect of the resident are affected. Therefore, the development of a dynamic sensing method capable of sensing the thermal state of a human body in real time becomes an important direction of building intellectualization and green energy saving research. In the prior study, as disclosed in CN112032971B, an indoor thermal environment regulation and control method based on heart rate monitoring is disclosed, and the main flow comprises the steps of firstly collecting indoor air temperature and humidity, calculating air-water separation pressure, then estimating the real-time metabolism rate of a user based on resting heart rate and real-time average heart rate, and then calculating the thermal sensation index of the user based on the index to regulate and control an air-conditioning system. According to the method, individuation and dynamics of thermal environment regulation are achieved by combining human physiological information with environment parameters, indoor thermal comfort is improved to a certain extent, but heart rate is used as a single physiological index, the human thermal state is difficult to comprehensively reflect, emotion, movement or external interference influence is easy to occur, thermal sensation prediction is conducted only based on air temperature and humidity and metabolic rate, local skin temperature change and thermal distribution of a human body cannot be accurately reflected, delay or error exists in dynamic thermal environment prediction, and continuous and high-precision indoor thermal environment dynamic assessment is difficult to achieve. Aiming at the problem, the invention provides an artificial intelligence-based indoor thermal environment dynamic evaluation method and an artificial intelligence-based indoor thermal environment dynamic evaluation system, which realize the dynamic perception evaluation of the indoor thermal environment and are beneficial to promoting the cross fusion development of intelligent buildings and energy-saving air conditioners. Disclosure of Invention The invention provides an indoor thermal environment dynamic assessment method and system based on artificial intelligence, wherein an indoor infrared thermal imaging sequence is easily interfered by environmental factors such as wind speed, humidity and the like, so that pixel value deviation is caused, skin temperature characteristic extraction and thermal comfort assessment are affected, S2, a pixel-level environmental temperature compensation item is adaptively generated by combining a low-frequency effective component of an environmental thermal sensing data sequence through a temperature compensation algorithm based on time-frequency domain joint analysis, a pure infrared thermal imaging sequence is obtained, environmental temperature time sequence characteristics are extracted, accurate input is provided for dynamic assessment, S3, multi-scale skin temperature decomposition and skin density we