CN-121048827-B - Radiation protection wall air tightness detection method based on image fusion

CN121048827BCN 121048827 BCN121048827 BCN 121048827BCN-121048827-B

Abstract

The invention relates to the technical field of air tightness detection, in particular to an image fusion-based radiation protection wall air tightness detection method which comprises the following steps of acquiring a protection wall multiband image through multispectral imaging, acquiring temperature and humidity, extracting wavelet transformation characteristics, constructing a time sequence set, calculating a change rate differentially, detecting an extremum through a sliding window, outputting a change rate vector, clustering Euclidean distance and included angle of space adjacent points, judging an abnormal area, modeling a polynomial trend line, and judging and outputting a positioning and tracking result. According to the invention, the temperature and humidity data of the whole area of the wall body are extracted in real time by fusing multispectral image information, the change rate analysis and trend line modeling are carried out by combining the sequence change trend and the spatial distribution characteristics, the suspected airtight abnormal area is accurately and efficiently identified and positioned, the change of the airtight performance is dynamically monitored and continuously tracked, the detection sensitivity and reliability are improved, the omission and erroneous judgment are reduced, the detection accuracy is ensured under the complex environment, and the comprehensive evaluation of the wall body airtight and early risk early warning are effectively supported.

Inventors

JIA SHIGUI
WU CHENG
LIU MINGFENG
LUO LIN
YANG TAO
WANG JIANDING
ZENG SHAOQIN
HUANG JIANG
SI WENJUN
WANG XIAO
ZHANG HAILIN
LIU YANSHENG

Assignees

中铁城建集团南昌建设有限公司
中铁城建集团有限公司
四川港通医疗设备集团股份有限公司

Dates

Publication Date: 20260508
Application Date: 20250829

Claims (10)

1. The radiation protection wall air tightness detection method based on image fusion is characterized by comprising the following steps of: S1, acquiring a full-area multiband image of a radiation protection wall through a multispectral imaging device, acquiring temperature values and humidity values of each detection point, inputting parameters into a wavelet transformation algorithm to perform feature extraction, and constructing a temperature-humidity feature time sequence set of the radiation protection wall; S2, acquiring continuous time sequence temperature parameter and humidity parameter changes of each detection point by calling the temperature and humidity characteristic time sequence set of the radiation protection wall and adopting differential calculation to construct a two-dimensional change rate vector, inputting the vector into a sliding window extremum detection algorithm for screening, and outputting a temperature and humidity change rate vector sequence of the detection point of the radiation protection wall; S3, judging the temperature and humidity change rate vector of the space adjacent detection points by applying Euclidean distance and vector included angle by calling the temperature and humidity change rate vector sequence of the detection points of the radiation protection wall, clustering based on synchronous trend, and outputting a detection result of a suspected airtight abnormal region; And S4, based on the detection result of the suspected airtight abnormal region, collecting a continuous time sequence temperature and humidity change rate vector of a detection point of the region, calling polynomial trend line fitting to model a vector trend line, inputting a trend line slope and fitting residual error into a trend line consistency judging model, and outputting the airtight abnormal positioning and dynamic tracking result of the radiation protection wall.
2. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 1, wherein the radiation protection wall temperature-humidity characteristic time sequence set comprises temperature distribution parameters, humidity distribution parameters and time sequence indexes, the radiation protection wall detection point temperature-humidity change rate vector sequence comprises a temperature change rate vector, a humidity change rate vector and a time sequence change label, the suspected air tightness abnormal region detection result comprises detection point space coordinates, a change rate synchronous group and an abnormal clustering label, and the radiation protection wall air tightness abnormal positioning and dynamic tracking result comprises air tightness abnormal detection point positions, dynamic track parameters and trend line fitting parameters.
3. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 2, wherein the step S1 comprises the following steps: Acquiring an infrared band image and a visible band image of the radiation protection wall in a full area, which are continuously acquired by the multispectral imaging device in a preset time interval, and synchronously monitoring real-time temperature values and real-time humidity values of all preset detection points on the surface of the radiation protection wall to generate an original multi-mode data set; Aiming at the real-time temperature value and real-time humidity value sequences in the original multi-mode data set, calculating the approximate coefficients and detail coefficients of each sequence under different decomposition levels by using a multi-scale wavelet decomposition algorithm to obtain a multi-scale characteristic coefficient set; calculating the energy of each level approximation coefficient and the energy of detail coefficient in the multi-scale characteristic coefficient set, taking the approximation coefficient energy as the temperature distribution parameter and the humidity distribution parameter, and jointly establishing the radiation protection wall temperature and humidity characteristic time sequence set by combining the time sequence information of the multi-band image.
4. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 2, wherein the step S2 comprises the following steps: Invoking the temperature and humidity characteristic time sequence set of the radiation protection wall, and extracting the temperature distribution parameters and the humidity distribution parameters at two continuous time points; Calculating the difference between the temperature distribution parameter of the later time point and the temperature distribution parameter of the former time point to obtain a temperature variation difference value, calculating the difference between the humidity distribution parameter of the later time point and the humidity distribution parameter of the former time point to obtain a humidity variation difference value, and combining the temperature variation difference value and the humidity variation difference value into the two-dimensional variation rate vector; Setting a sliding window size and an extremum judging threshold value, moving the sliding window on a time sequence formed by the two-dimensional change rate vectors, judging whether the maximum value of vector modes in the window exceeds the extremum judging threshold value, screening out the two-dimensional change rate vectors exceeding the threshold value, and obtaining a temperature and humidity change rate vector sequence of the radiation protection wall detection point.
5. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 2, wherein the step S3 comprises the following steps: Invoking the temperature and humidity change rate vector sequence of the detection points of the radiation protection wall, and determining a space adjacent detection point set of each detection point according to the space coordinates of the detection points; Calculating Euclidean distance and vector cosine similarity between the temperature and humidity change rate vector of the target detection point and the temperature and humidity change rate vector of each detection point in the space adjacent detection point set, and generating a multidimensional correlation characteristic parameter; Setting a distance upper limit threshold and a similarity lower limit threshold, judging whether the Euclidean distance is smaller than the distance upper limit threshold and whether the vector cosine similarity is larger than the similarity lower limit threshold, and judging detection points meeting the conditions as synchronous change point pairs; And performing aggregation analysis on all the synchronous change point pairs by applying a density clustering algorithm, marking a detection point set with the same spatial communication and synchronous change trend as the change rate synchronous group, and generating the detection result of the suspected airtight abnormal region.
6. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 2, wherein the step S4 comprises the following steps: Extracting temperature and humidity change rate vector sequences of all detection points in the change rate synchronous group in the radiation protection wall detection points in a continuous time period according to the detection result of the suspected airtight abnormal region; For the temperature and humidity change rate vector sequence of each detection point, performing polynomial fitting by adopting a least square method, establishing a unique polynomial trend line for the temperature and humidity change trend of each detection point, and calculating to obtain the slope of the trend line and the fitting residual error; Analyzing the deviation degree of the slope of the trend line and the dispersion degree of the fitting residual error based on a preset reference slope and a residual error threshold, and judging that the detection point is an airtight abnormal point when the deviation degree of the slope and the dispersion degree of the residual error exceed the preset threshold; and integrating the position information of all the airtight abnormal points, recording the position change in the continuous time, and generating the airtight abnormal positioning and dynamic tracking results of the radiation protection wall.
7. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 5, wherein the judgment of the synchronous trend is realized by calculating a synchronous trend judgment index Completing; The synchronous trend discrimination index The calculation formula of (2) is as follows: ; Wherein, the For the target detection point Detection point adjacent to space Judging indexes of the synchronous trend; As the euclidean distance weighting coefficient, Is the cosine value weight coefficient of the vector included angle and ; Is the detection point Temperature-humidity change rate vector of (2) And detecting point Temperature-humidity change rate vector of (2) Euclidean distance between them; The maximum normalized distance is preset; Calculating the synchronous trend discrimination index Then comparing the synchronous trend with a preset synchronous trend judging threshold value, if the synchronous trend judging threshold value is the same as the synchronous trend judging threshold value And if the detection point is larger than the synchronous trend judging threshold value, judging that the two detection points have synchronous trend.
8. The method for detecting the air tightness of the radiation protection wall based on the image fusion according to claim 6, wherein the trend line consistency judging model is used for calculating a trend line deviation index Realizing; The trend line deviation index The calculation formula of (2) is as follows: ; Wherein, the Is the first Trend line deviation indexes of the suspected airtight abnormal area detection points; as a weighting factor for the trend line relative slope deviation term, A weighting factor for normalizing the fitting residual root mean square error term; Is the first Fitting a slope to the polynomial trend line of each suspected airtight abnormal region detection point; is the slope of a reference trend line of temperature and humidity change of the radiation protection wall under the normal airtight condition, and the absolute value thereof Is not zero; Is the first The detection points are at the time points The difference between the actual temperature and humidity change rate vector and the fitting value of the polynomial trend line is the fitting residual; the total number of the continuous time sequence temperature and humidity change rate vectors used for fitting; fitting the root mean square error value of the residual error for the reference under the normal air tightness condition, and Is not zero; calculating the deviation index of the trend line Then comparing it with preset trend line consistency threshold value, if And if the detected point is larger than the trend line consistency threshold value, finally confirming that the detected point is the airtight abnormal detected point.
9. The method for detecting the air tightness of the radiation protection wall based on image fusion according to claim 2, wherein the sliding window extremum detection algorithm is specifically implemented as follows: Setting the size of the sliding window as an integer greater than 2, wherein the integer is the length of a time sequence segment for extremum detection; setting the moving step length of the sliding window as a positive integer, wherein the positive integer is a parameter for controlling the overlapping degree of two adjacent detections; For a complete time sequence formed by the two-dimensional change rate vectors of each detection point, starting from an initial moment, successively intercepting a vector subsequence with the length equal to the sliding window size according to the moving step length; inside each vector subsequence, calculating Euclidean norms of all two-dimensional change rate vectors, and identifying a vector with the maximum norm value as a local extremum vector of a current window; comparing the Euclidean norm of the local extremum vector with a preset extremum judging threshold value, wherein the extremum judging threshold value is a normal temperature and humidity fluctuation upper limit obtained based on historical data statistics; if the Euclidean norm of the local extremum vector is greater than the extremum judging threshold, reserving the vector and adding the time sequence change label to the vector, and finally, all reserved vectors jointly form the temperature and humidity change rate vector sequence of the radiation protection wall detection point.
10. The method for detecting the air tightness of the radiation protection wall based on image fusion according to claim 3, wherein the calculating of the approximate coefficient and the detail coefficient of each sequence under different decomposition levels by applying a multi-scale wavelet decomposition algorithm is specifically as follows: selecting a Daubechies wavelet basis function to carry out multi-layer decomposition on the real-time temperature value sequence and the real-time humidity value sequence of each detection point, wherein the number of decomposition layers is determined in a self-adaptive manner or preset according to the non-stationary characteristic of the signal; at each decomposition level, decomposing the signal into said approximation coefficients representing low frequency trends of the signal and said detail coefficients representing abrupt changes in the signal; Extracting the approximation coefficient of the last layer of decomposition; Extracting the detail coefficients of all decomposition levels; And forming a characteristic vector by the extracted approximation coefficients and the detail coefficients of all levels together, and generating a multi-scale temperature-humidity dynamic characteristic vector of the detection point.

Description

Radiation protection wall air tightness detection method based on image fusion Technical Field The invention relates to the technical field of air tightness detection, in particular to a radiation protection wall air tightness detection method based on image fusion. Background The technology for detecting the air tightness of the radiation protection wall belongs to the technical field of detection and measurement, and has the core application scene of detecting and scientifically evaluating the airtight performance unfolding system of the radiation protection wall in the service period of the radiation protection wall. The technical field has clear and definite core working scope, and mainly comprises five key matters, namely, structural integrity detection of a radiation protection wall, ensuring that a wall foundation structure has no damage or defect affecting tightness, air tightness parameter measurement, accurate acquisition of core data reflecting the sealing degree of the wall, leakage point identification, positioning of a possible gas leakage position of the wall, detection of environmental pressure control, construction of a stable detection environment to ensure data accuracy, and sealing performance evaluation, and grade judgment and risk evaluation of the sealing effect of the wall by integrating various data. From the technical coverage, the application of a physical detection method and professional detection equipment is integrated in the field, and a set of relatively complete technical system is formed through the whole process links of detection data acquisition, parameter depth analysis, detection standard formulation and the like. In a radiation protection wall air tightness detection technology system, a traditional detection algorithm is an early-applied and basic detection method. The method is characterized in that the gas pressure with a specific value is applied to the radiation protection wall, then the change condition of the pressure inside the wall is monitored in real time by means of a professional pressure change measuring instrument, and finally the sealing performance of the wall is evaluated according to pressure change data. In the actual operation process, the traditional detection algorithm generally adopts two pressure application modes, namely, one pressure application mode is used for applying positive pressure on one side of the wall body, and the other pressure application mode is used for applying negative pressure. In any mode, the pressure attenuation value in the wall body in unit time can be continuously recorded by matching with high-precision instruments such as a digital pressure gauge, a micro differential pressure gauge and the like, or the change condition of the gas flow is calculated through a related device. The detection personnel can analyze the pressure drop rate according to the real-time monitoring data or directly calculate the gas leakage flow rate, so as to be used as a key basis for judging the gas tightness level of the wall body, namely, if the pressure attenuation is slow and the gas leakage flow rate is small, the gas tightness of the wall body is good, otherwise, the wall body is indicated to have sealing defects. However, in the practical application process, the conventional airtight detection method has obvious limitations, and the limitations affect the accuracy and reliability of the detection result to a certain extent. The method has the core problems that the traditional detection mode only depends on two parameters, namely pressure change and gas flow, to judge the air tightness of the wall body, and the consideration of other factors possibly influencing the detection result is insufficient. Specifically, in an actual detection environment, interference of an external environment (such as irregular flow of air flow in a detection field, sudden fluctuation of external air pressure, etc.), a complex structure of a radiation protection wall (such as a splicing structure of a wall body with multiple layers of different materials, penetration of a pipeline in the interior, etc.), and abnormal temperature and humidity of a local part of a detection area may generate interference on detection data. For example, when unexpected fluctuation occurs in pressure in the detection environment, deviation can be directly caused in pressure attenuation data, so that it is difficult for detection personnel to accurately judge whether pressure change is caused by wall leakage or ambient pressure fluctuation, when the wall structure is uneven or a multilayer structure exists and local temperature difference is obvious, the flowing state of gas in the wall becomes complex, at the moment, a tiny leakage phenomenon cannot be accurately distinguished only through pressure measurement, namely, the pressure change caused by temperature difference can be misjudged as leakage, the pressure change caused by tiny leakage points can be omitted, finally,